0:00

On this episode of data-driven, we

0:02

chat with Istven Mizoros, the founder

0:04

of Mitsu, a startup on a

0:06

mission to make warehouse native analytics

0:08

not just powerful, but downright delightful.

0:11

We dive into how Mitsu empowers startups

0:13

and enterprises alike, what it's like

0:15

to build a data company in

0:17

Europe, and yes, the mysterious law

0:19

of their hello kitty adjacent branding.

0:22

Istven shares his journey from

0:24

CERN physicists to open source

0:26

evangelist to startup founder. A

0:28

quick heads up. There were

0:30

the occasional glitches in our

0:33

guest's audio. Frank did

0:35

his best to clean it

0:37

up, but alas, his only

0:40

human. Clearly, he should

0:42

have had an AI

0:44

do it. But the

0:46

conversation is still well worth

0:49

a listen, full of insights,

0:51

data, and a healthy respect for

0:53

the long sale cycle. data science

0:55

and of course it's all back-ended

0:57

by data engineering and data engineers.

0:59

My favorite is data engineer in

1:01

a world will not be able

1:03

to make it today but he

1:06

will be here in spirit and

1:08

I'm sure we can if we

1:10

have a good conversation and he

1:12

will definitely want to talk to

1:14

you. But today I have with

1:16

me Istavan Miss Sarios and I'm

1:18

not sure if I pronounce that

1:20

right. He is the... It's good

1:23

enough. All right. Yeah. He is

1:25

the co-founder of meetsu.io and it

1:27

is a platform that empowers warehouse

1:30

native decision-making for startups and

1:32

enterprises alike without compromising data

1:34

quality security or performance. Welcome

1:37

to the show. Yistan. Thank

1:39

you. It's good to be

1:41

here. Yeah. Yeah. So we were talking

1:43

in the virtual green room that the

1:46

name of the company is this meets

1:48

Cat. Like and you have a whole

1:50

lore behind the. Yeah, there is multiple

1:52

stories. It's a name of a

1:54

cat in my native tongue. We

1:57

used to have this cat

1:59

called Mitsu. Oh yeah. And it's just

2:01

a digital letter in the end,

2:03

it's too. And the other story

2:05

is, where it came from, is

2:07

that there is a restaurant which

2:10

is very similar to named in

2:12

the city of Ryan's, like Budapest,

2:14

Hungary, and with one of my

2:16

early partners in this business, who

2:18

is not anymore with us, we

2:20

were brainstorming out the name, and it

2:23

took like two three hours until like

2:25

so pissed, then we said like,

2:27

let's just call it, me too.

2:29

It's a restaurant restaurant called Mitzi.

2:32

That's funny. And just change the

2:34

name again, like the letter, the

2:36

last letter, and it became the

2:38

name of the company. That's cool.

2:40

Yeah, it tries to kind of appeal

2:43

to the, to like

2:45

developers, like like individual

2:47

contributors. So it's like a kind

2:49

of cool tool that you could like,

2:51

I mean, and start using. Originally

2:54

it was open source. So it was also

2:56

kind of going through that direction even further,

2:58

like we want to make it stick to

3:00

the development community, to the data

3:03

engineering community. So it's like easy to

3:05

remember, it's friendly. It's not an enterprise

3:07

solution. Originally, it wasn't an enterprise at

3:10

all. We didn't defend the present in

3:12

our mind. So it tried to be

3:14

appealing to the open source community. It's

3:16

too late now to change it. I

3:18

stuck with us. I like it actually.

3:21

It's very, it's very reminiscent and the

3:23

logo looks like a little like cat.

3:25

So, kind of with a, with just

3:28

a little bit of that reminder

3:30

of hello kitty, but not enough

3:32

to get the lawyers mad at

3:34

you. Which is important to not

3:36

get lawyers mad at you. But

3:38

my listeners have heard all about

3:41

my custody fight to get the

3:43

littlest guy, but that's. So what's interesting

3:45

is you were talking about the lore we

3:47

were in the virtual agreement we're talking about

3:49

lore around branding and I think that is

3:51

the kind of the new I don't know

3:54

it seems like a new trend like you look at

3:56

do a lingo do a lingo you know they don't

3:58

just have they used to have the owl and

4:00

then the owl kind of

4:02

became this monstery type thing.

4:04

And now there's a whole

4:06

slew of characters with it. And

4:09

my YouTube feed is filled with

4:11

it. There's Lily, there's some

4:13

other girl, there's some

4:15

guy and his son, like

4:17

there's this whole community around

4:20

this. And I think in

4:22

terms of branding, I think

4:24

that's really the next level

4:26

of. You know, which makes me think

4:29

because we have Bailey our virtual

4:31

kind of master ceremonies or mistress

4:33

of ceremonies. I'm not sure how

4:36

that goes virtual MC and You

4:38

know, she definitely has a law that we

4:40

never really fully published but You

4:42

know, it's kind of like it

4:44

helps inform kind of creating your

4:46

character and when I have an

4:48

I created like a custom GPT

4:51

that will talk in her style

4:53

and kind of have her kind

4:55

of So I think that's

4:57

interesting. So you started

4:59

off as open source as a

5:02

lot of startups do. And

5:04

in my day job I

5:06

work at Red Hat. So

5:08

obviously they're big in open

5:10

source. Are you no longer open

5:13

source? Yes, yes. We about

5:15

done it. Last we got

5:17

the funding from VCs. It's

5:20

an uphill battle. It's

5:22

I would say twice as hard

5:24

or maybe more times like harder

5:26

to do an open source startup

5:28

especially from Europe. Okay. It's it's just

5:30

it's just an up a dot. And

5:33

we've seen with this open source

5:35

initial detection in the in

5:37

the local you know like by the

5:39

way before we get started with

5:41

that like it's a be to be

5:43

solution in the world. What you

5:45

are building is a be to be

5:48

solution. So we started to see

5:50

initial attraction with the open source version

5:52

of the application in the local

5:54

market essentially startups companies started to use

5:57

it in the open source version and

5:59

that's what got us funded. But

6:01

again, if you if you rent

6:03

is open source, you self host

6:05

it or something, it's very hard

6:07

to monetize it. You know,

6:10

so we decided to shut it

6:12

down and essentially make it like

6:14

a SAS application, what we're building.

6:16

And that's what we can. And

6:18

the brand stuck with us, I

6:21

believe in. You know, you mentioned Doling,

6:23

it's a great, great example. I

6:25

love their brand so much. It's

6:27

amazing. I love the application. I

6:29

love everything about Doling, actually. It's

6:31

a fantastic what they are building.

6:33

The difference is they are B to C. Yeah,

6:35

yeah, I mean, that makes sense. Yeah, customer's

6:38

example. We are keeping these, you know,

6:40

to kind of stand out from the

6:42

crowd. Our current customer base is

6:44

in the upper market enterprise segment,

6:47

but I still believe we are selling

6:49

the humans, you know, you know, you know,

6:51

after all, There must be some type

6:53

of branding that then remember. It

6:55

shouldn't be this dark blue, I

6:57

don't know, boring colors, or the other

7:00

color, current coloring is a bit too

7:02

wild for it, I'm pretty sure about

7:04

that. But all the competitors we have

7:06

is like is boring, you know,

7:09

no branded, like weird logos, some

7:11

abstract stuff that you don't

7:13

remember. So with a brand we are

7:15

trying to stand out in the B2B, you

7:18

know, segment still, that we are

7:20

kind of memorable. With that said,

7:22

we are, we should put a little, so

7:24

much more effort in this brand, you know,

7:26

like, you know, emails, posting about us,

7:28

should have branding, but we kind of don't

7:30

do it. We don't have the time, don't

7:33

have the resources for it, but

7:35

ultimately we should, we should definitely

7:37

do. My inspiration is a service called

7:40

Vanta. Maybe you have heard about them.

7:42

It's there like security. No, tell

7:44

me more. So they are like, they are

7:46

not in the data space, they

7:48

are in the security and compliance

7:50

and you know, they do for businesses,

7:53

SOC 2 or ISO, all these

7:55

like kind of compliancy, you

7:57

know, auditing and all that is around

7:59

that. It's a super, like it cannot

8:01

be more be-to-be than that. You know, it's

8:03

like only enterprises or like, like,

8:05

you know, growing startups need that

8:08

solution. But their brand is still,

8:10

somebody resembles this lava kind

8:12

of animal that they started

8:14

with originally. And I like that

8:16

because now I still remember them and I'm

8:18

telling you about this, you know, and this

8:20

is what it means, what it means, the

8:23

brand. So we can have a conversation about it

8:25

without even, you know, they are trying to push their narrative

8:27

to us. Just because the brand is there, you know, they

8:29

have a, it's, it's, it's, it's, it's, it's, it's, it's, it's,

8:31

it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's,

8:33

it's, it's disappearing from the website, from the website, you

8:35

know, I think you're right, you're right, I think you're

8:37

right, I think you're right, I think you're right, I think

8:40

you're right, I think you're right, I think you're right, I

8:42

think you're right, I think you're right, I think you're

8:44

right, I think you're right, I think you're right, I think

8:46

you're right, I think you're right, I think you're right, I

8:48

think you're right, I think you're, I think you're until

8:51

the AI agents take over, will

8:53

always be selling to humans. And

8:55

branding is one of those things

8:57

that's funny, right? So now I

9:00

find myself in my day job

9:02

in a marketing capacity, which is

9:04

something I never would have envisioned

9:07

in my life at all. But there

9:09

is a need even in highly

9:11

technical marketing space to be memorable,

9:14

right? Some would argue even more

9:16

so. There's always trends in marketing,

9:18

right? Because marketing is not in

9:21

isolation, right? Because the colorful logos,

9:23

you mentioned the colorful logos,

9:25

versus the boring logos. You are

9:27

competing for mindspace with. Yes. Even though

9:29

you're strictly speaking, B to B, right? And

9:32

do a lingo as B to C. But

9:34

you are competing for mind space with that

9:36

green owl, right? Yes. Not so much directly,

9:38

but like, you know, there's only so much.

9:40

human minds can have in their head at

9:43

the same time. And I highly recommend viewers

9:45

to watch one of my most favorite videos

9:47

of all time. It used to be on

9:49

Channel 9, which was a Microsoft property. That

9:51

has since been doing, but just do

9:54

a Google or YouTube search for human

9:56

operating system. Bill Hill, human operating system.

9:58

And he basically talked. about the gist

10:00

of it is only a three minute video,

10:03

but I saw it in 2004 or something

10:05

like that and it stuck with me how

10:07

many years later, 20 years later. And

10:09

it's basically the idea is that

10:11

the most important operating system in

10:13

the world is in iOS, it's

10:15

on Android, it's on Windows, it's

10:17

the human operating system and it

10:19

was built for different circumstances, there's

10:21

no update in sight, and you

10:23

kind of have to deal with

10:25

the quirks of that. And I

10:27

think if you come from branding,

10:29

particularly as a technical geek, like

10:31

you or me, where we have to think

10:33

in those terms, it's helpful because branded

10:35

for a lot of technical people is

10:37

a struggle. But if you think about,

10:39

you know, you're not, you're basically writing

10:42

code or building experiences for the human

10:44

operating system, and that's a good way

10:46

to do it. So I actually kind

10:48

of like the logo. I don't think

10:50

it's, I think. Yeah, I guess for

10:52

a B to B kind of thing,

10:54

it's a little flashier than normal, but

10:56

I still when I look at that

10:58

thing, I'm reminded that Hello Kitty and

11:00

I'm sure there's a lot of Hello

11:02

Kitty fans out there and that and or

11:04

have, you know, relatives or kids that

11:06

that are into that. So yeah, yeah.

11:08

So go ahead. Sorry, I like to look

11:11

at allergies. It's like, um, and we did

11:13

the first got the funding. Many

11:15

VCs came back to us like a

11:17

package, locat and branding. It basically used

11:19

this word. It's like it needs to

11:22

work. It's tries to, I'm trying to

11:24

remember that this was a. Yeah, I need to

11:26

explain it a lot of times.

11:28

So, but it's still, it's still,

11:30

I think it works. It works,

11:32

you know, and that's just it.

11:34

Like, that's what they remember, right?

11:36

So even if they don't, and

11:39

I don't want to speak for

11:41

VCs, I have a lot of

11:43

friends or VCs, but they have

11:45

a lot of people coming at

11:47

them, right? So like, if you

11:49

can kind of plant a little

11:51

seed in their head, it's like,

11:53

what was the name of that

11:56

data company that, you know, the

11:58

Hello Kitty logo, right? by

12:00

lawyers so it doesn't it's not

12:02

exactly a copy of Hello Kitty

12:04

but but but yeah so like

12:06

what was the what was the

12:09

gap in the market that made

12:11

you go after building this because

12:13

there's a big a lot of

12:15

big players in there and I

12:17

see you integrate with all you

12:19

integrate with all of them snowflake

12:22

redshift data bricks and there's a

12:24

handful of others so what what

12:26

was the gap in the market

12:28

that you yeah you decided to

12:30

fill Maybe the best way to

12:32

explain it. This is my story.

12:35

The problem that I found, actually,

12:37

I was facing that problem. I

12:39

was working in the BTC segment

12:41

as a data engineer, dimly sometimes

12:43

for data engineering, in various B2C

12:45

companies, e-commerce, travel, sometimes like SAS,

12:48

but it's like low-deal size like

12:50

SAS. So means like B2C SAS,

12:52

essentially. $20 deals, $10, $10 deals.

12:54

So the issue there is that

12:56

these companies, and I would actually

12:59

highlight all this travel industry, because

13:01

the problem is I believe is

13:03

the most visible there, that you

13:05

have so much data, like you

13:07

are optimizing for the volume of

13:09

data, you must optimize, maximize the

13:12

number of visitors on your website.

13:14

Imagine kayak or like Google Search

13:16

or any other search, flight search

13:18

providers like any type of like

13:20

travel industry, search provider, trip advisor

13:22

for example, their business in the

13:25

number of visitors, you know, so

13:27

naturally you must maximize the number

13:29

of visitors in your site, therefore

13:31

you are basically generating a huge

13:33

volume of data. That's one thing.

13:35

The other thing that they need

13:38

to. the optimize for is the

13:40

is the conversion rates on the

13:42

on the on the on the

13:44

website how many people buy actually

13:46

a flight or like they they

13:49

search for a trip or something

13:51

like that so you have on

13:53

one hand you have this which

13:55

amount of data in the visitor

13:57

space. And the other hand, you

13:59

or the marketing people and your

14:02

product people must maximize the conversion

14:04

rates. So there is a need

14:06

for a product and marketing analytics

14:08

tool by these people, credit market

14:10

managers, but there is also a

14:12

huge amount of data. So current

14:15

solutions in this space, I can

14:17

name a couple of brands, but

14:19

I don't want to name brands,

14:21

but there are a couple of

14:23

solutions that are the increments in

14:25

this space. They are not very

14:28

well suited for huge volumes of

14:30

data. Because the current model they

14:32

operate on is actually based on

14:34

the volume of data. They charge

14:36

based on the volume of data.

14:38

Which is for a travel industry

14:41

company, it's a kind of problem

14:43

because they might get more data,

14:45

but it doesn't mean more revenue

14:47

immediately for them. Right.

14:49

I'm mainly explaining a bit too long

14:51

this, but this is where the problem

14:54

comes in. So there is a segment

14:56

of companies that want to do product

14:58

and marketing analytics with their own data

15:00

because everything else became unfeasible for them.

15:03

Okay. So I give an example. Yeah.

15:05

So I give an example like, let's

15:07

say a company kayak and just give

15:10

a given example. They have monthly 100

15:12

million visitors visitors. You know. If you

15:14

would send that amount of data to

15:16

the current income and product and marketing

15:19

and lighting solutions, you would pay millions

15:21

of dollars yearly just to measure the

15:23

commer generate. Right, because there's also the

15:25

web analytics like mouse clicks. Yes, you

15:28

know, where their cursor lingered time on

15:30

the site, like there's all sorts of

15:32

data there and they, you know, as

15:34

someone, you know, Franks World.com is my

15:37

site, right. I don't get millions of

15:39

visitors, but I get about, you know,

15:41

you know, 200,000 page years a month

15:43

and I've looked at these analytic solutions

15:46

and I see the bill and I'm

15:48

like look I would love to know

15:50

all this but I don't have the

15:52

budget for that right so I can

15:55

only imagine what some like a kayak

15:57

or a Google would have, right? Like,

15:59

and the amounts of data that they

16:01

have are probably just ridiculously large in

16:04

the cost of doing it. Probably the

16:06

juice isn't worth a squeeze as of

16:08

now, as we say in the US,

16:10

right? Like, you know. Yeah, exactly. So

16:13

you have like maybe 20 people in

16:15

marketing, but you have daily, mainly one

16:17

terabyte of data, right. You would pay

16:20

10x the amount of money. the corporate

16:22

provider than to your employees, essentially. That

16:24

was the mass that we had in

16:26

my previous experiences. And on top of

16:29

that, so you, by the way, we

16:31

need to mention that you need as

16:33

a company a product and marketing analytics

16:35

tool that is self-service. That's the goal

16:38

here, to have a self-service capability for

16:40

the management for the product and marketing

16:42

people to analyze their data on their

16:44

own. Otherwise, you need to hire a

16:47

lot of data analysts. That is, again,

16:49

extremely costly, very slow. It's just a

16:51

lot of risk coming with it, just

16:53

to build a data team you need

16:56

to make management and everything. So you

16:58

need this self-service tool to analyze conversion

17:00

rates in your application, but you have

17:02

too much data. And essentially this was

17:05

kind of like the boss, the original

17:07

investment didn't turn out to be actually

17:09

feasible to use any of the third-party

17:11

tools. And on top of that, the

17:14

company was working for, they had their

17:16

own data, warehouse data, warehouse data, or

17:18

data lake, let's say, where they collected

17:20

the same data. And that data was

17:23

anyway already present in the company. So

17:25

the idea was, why don't we build

17:27

a product analytics tool on top of

17:30

the database? So basically. Yeah, that makes

17:32

sense. Yeah, because like the date, like

17:34

that's what people say little, you know,

17:36

data processing is cheap and data storage

17:39

has never been cheaper. Yeah. But the

17:41

analytics tools, you know what? That's a

17:43

little bit of a different thing. So

17:45

I think it sounds like from what

17:48

I can kind of determine that you

17:50

built a solution that is kind of

17:52

a for lack of a term a

17:54

web wrapper around existing products because I

17:57

I deal with this a lot in

17:59

my current job and previous jobs is

18:01

that once a data warehousing, you know,

18:03

or our DBMS solution has been in,

18:06

it's in there for decades, right? Because

18:08

no one wants to take and rip

18:10

everything up. And, you know, data breaks

18:12

is a good example, right? Like they,

18:15

once they're in a company there, they're

18:17

in there, right? That's why you have.

18:19

you know all the cloud providers are

18:21

rushing to to add support for them

18:24

right this way I was there when

18:26

they added Azure support and you know

18:28

at Microsoft and like hey why are

18:30

we doing this because they're a big

18:33

player right and Microsoft was one and

18:35

that was one of those decisions that

18:37

Microsoft is one of the few companies

18:40

in the world that could build a

18:42

platform that rivals. data breaks. Some people

18:44

say that they did with fabric and

18:46

synapse. I'm not going to open that

18:49

can of worms because Andy's not here

18:51

to speak on that. But let's just

18:53

say, like they have, they had the

18:55

means, right? Whether or not they did

18:58

it, let's put that conversation aside. But

19:00

they have the means to put that

19:02

conversation aside. But they have the means

19:04

to do it. But what did they

19:07

decide to support data breaks because I

19:09

think that you're on to something that

19:11

once once once these solutions are in

19:13

place these solutions are in place, they're

19:16

in place. No, this is short just

19:18

to like I was talking about this

19:20

like whole problem for like a minute

19:22

or two and it's a written or

19:25

it's better to like a one like

19:27

one line pitch is basically it's a

19:29

product analytics tool on top of your

19:31

data warehouse. What it does essentially is

19:34

just automatically generate sequel queries for like

19:36

funnel conversion rates, retention rates measurement journey

19:38

type of like visualization imaging like a

19:41

same key chart. You get like number

19:43

of unique user sessions. All these kinds

19:45

of questions that a marketing person or

19:47

like a product person would ask on

19:50

a daily basis, which are multiple times,

19:52

with those questions you don't have to

19:54

go to a data analyst, you can

19:56

just go to the application, it will

19:59

answer automatically from you, from Databricks, from

20:01

Snowflake, from Big Prairie, from all these

20:03

data warehouses, with Bayby. automatically generating sequel

20:05

queries. This is what we do. This

20:08

is what we do. The problem was

20:10

basically too much data. People cannot access

20:12

the data without the data list, but

20:14

they cannot use the third party tool

20:17

because this is basically, they cannot pay

20:19

it. It's just too much. Multiple millions

20:21

of dollars per year. So that was

20:23

the original problem. So it's quite simply

20:26

simply explained, like, there are certain companies

20:28

that they cannot, cannot afford the incumbent

20:30

solutions. or product Olympics. Well, exactly, like,

20:32

you know, I'm, I'm a, I'm a

20:35

basically a one, two person operation, maybe

20:37

three, like, you know, it's like, and,

20:39

you know, our monetization, we are working

20:41

on that, like, you know, our monetization

20:44

strategy, but, you know, part of our

20:46

success is we keep production costs low.

20:48

Yeah. And our reach high, right. So

20:51

like, because I don't, you know, I

20:53

leverage AI for a lot of things

20:55

that to make us look bigger, like,

20:57

like, like, like, like, like, like, you

21:00

know, you know, I'm sure your team

21:02

would be interested in this solution. And

21:04

I'm like, oh, that's cute. They think

21:06

I have a team. But I should

21:09

get my teenager this summer, like, when

21:11

he's on break, like doing some analytics

21:13

work for me. So, but no, so

21:15

I think you're right, because I'm looking

21:18

at the industries that you work with,

21:20

right? Obviously, we mentioned e-commerce. But also

21:22

you have media and entertainment and mobile

21:24

games. And those are both fields with

21:27

pretty thin margins, I would say, right?

21:29

In B2B SAS, I would imagine also,

21:31

I wouldn't say they have thin margins,

21:33

but they probably, particularly as they're starting

21:36

out, they probably have a lot of

21:38

challenges with funding and stuff like that.

21:40

So, B2C says is like a very

21:42

interesting domain. B2B, you know, when you

21:45

have like 10 clients and you survive

21:47

that, like on the kinds, in some,

21:49

we have some customers like B2 Interesting.

21:52

This mostly B to C you say.

21:54

Yes. The B to be segment for

21:57

us. is interesting for another reason, because

21:59

the solution we provide inherently is compliant

22:01

with their like data or requirements, it's

22:04

exactly required. The data never leaves your

22:06

data warehouse, to do analytics, you do

22:08

do self-service analytics, therefore it stays in

22:10

your data warehouse. We are just like

22:13

a BI2 essentially that attaches to that

22:15

data warehouse, but you don't have to

22:17

go through this all this kind of

22:20

like processes to validate the vendor like

22:22

us. because essentially we don't copy your

22:24

data. We don't get your raw data.

22:27

We just run a sequel query. We

22:29

get some results. We present them. And

22:31

that's it. So the data team doesn't

22:34

have to come. It's like a lot

22:36

of people internally. That is a good

22:38

one. And moreover, we also support on-prem.

22:40

Oh, really? OK. Interesting. And to be

22:43

fair, it's only AWS at the moment.

22:45

So if you're running AWS. we could

22:47

support the other ones as well but

22:50

it's like much more work. Well end

22:52

up less we have it off of

22:54

the box we can like deploy ourselves

22:57

onto your account and if you need

22:59

that extra extra you know security layer

23:01

that you are actually holding the service

23:04

as well then that we can provide

23:06

that. So that is one interesting topic

23:08

for B2B. The other segments B2C commerce

23:10

travel it is mostly about the volume

23:13

of data that just cannot move away.

23:15

Right. Yeah I mean that makes sense

23:17

like. Again, I'm a relatively small site

23:20

and I look at the pricing of

23:22

these tools and I'm not, it's way

23:24

out of my budget, right? Like, and

23:27

it's just, you know, so like, and

23:29

I always, I always do like these,

23:31

you know, I'm not your average Joe

23:33

or Jane, but like I can, I

23:36

do like I have my own LLLM

23:38

server, right, but like I also. You

23:40

know scale it down right because you

23:43

know I have videos on how I

23:45

built it, but basically it's a $300

23:47

repurposed Like VEL server from 10 years

23:50

ago, right? It's good enough for one

23:52

or two users, but I'm only ever

23:54

going to have one or two concurrent

23:57

users, right? But I can kind of

23:59

like, you know, do the math and

24:01

kind of, you know, explain what would

24:03

that look like in enterprise scale. It's

24:06

not perfect, but I do get a.

24:08

ballpark insight into how things would would

24:10

scale up. But what's interesting is that,

24:13

you know, you are based in the

24:15

EU. So I assume that you, and

24:17

even if you're not based in the

24:20

EU, you always think about GDPR and

24:22

things like that. So I guess there's

24:24

different classifications in GDPR, there's data processor

24:26

and data holder or something like that.

24:29

So what would you be in that

24:31

category? It's a good question. You had

24:33

a lawyer, a lot of lawyer. A

24:36

lot of lawyer. Good to be a

24:38

lawyer in the EU, I guess. We

24:40

have a data participant with companies. Yeah,

24:43

so we don't hold the data. We

24:45

hold like the data, which is our

24:47

analytics for the customer, the users. So

24:50

you have to give your email, of

24:52

course, and to restore the email address

24:54

of yours or your first name last

24:56

name, but that's it. But we don't

24:59

hold the customer data at all, essentially.

25:01

But that's a that's a big strategic

25:03

advantage because you there's a lot of

25:06

compliance things you can you're not really

25:08

when say you not your there's a

25:10

lot of regulations that won't apply to

25:13

you because you do it that way.

25:15

Oh yes yes yes yes yes and

25:17

that probably makes your sales cycle a

25:20

lot easier and your what say there

25:22

is definitely some education to this right

25:24

right right right right. I wouldn't say

25:26

it's easy to sell this. or yes

25:29

I would say the it's much you

25:31

know there in the sales there is

25:33

like a person can say no to

25:36

it but cannot say yes to it

25:38

is security and legal that can say

25:40

no to it and with the security

25:43

part we I mean didn't have a

25:45

no yet you know we had some

25:47

questions but is like you know even

25:49

without associate to or all this kind

25:52

of like compliance in whatever requirements we

25:54

were not you know kicked out of

25:56

the door so it's that's a good

25:59

one again that's cool like first case

26:01

we have to be self-hosted and then

26:03

this doesn't matter for the company like

26:06

if you if you're associate or not

26:08

right so like I like the fact

26:10

that you you say like down there

26:13

on your own page like no sales

26:15

calls no credit card needed so that's

26:17

cool so yeah that I may I

26:19

may poke around on it I have

26:22

I have data I Let's probably talk

26:24

to Andy because I have tons and

26:26

tons of data and you know I'm

26:29

a data scientist You know one-time data

26:31

engineer one-time software engineer and My shoe

26:33

you know shoemaker's children don't have good

26:36

shoes right like so I have I

26:38

I am like the worst I have

26:40

I have Excel files in various sharepoint

26:43

sites and in places and like where

26:45

I track all this stuff when every

26:47

time I go through it I'm like

26:49

I probably should do something with this

26:52

But yeah, so I definitely I'll definitely

26:54

it's it's got my got me curious.

26:56

So let's talk about how how did

26:59

like so you mentioned you were a

27:01

data engineer. And so there's a lot

27:03

of data engineers who listened to the

27:06

show like what was that moment where

27:08

you decided I want to start a

27:10

company. And like. I mean, I think

27:12

all of us have had those fantasies

27:15

like, I want to start a company,

27:17

but I know there's a lot of,

27:19

like, what was the moment that made

27:22

you like, like, do it? And you

27:24

also have young child. So like, it's

27:26

not like, you know, it's not like,

27:29

you know, you're the stereotypical founder with

27:31

no kids, no spouse, no nothing, right?

27:33

So what was that moment that made

27:36

you want to jump? Yeah, yeah, so.

27:38

I was like full-time employee for multiple,

27:40

multiple years and various companies, but at

27:42

one point I kind of like left

27:45

20, 40, so during COVID times. I

27:47

said I'm like, I'm doing the same.

27:49

thing for the fifth company now, you

27:52

know, like I'm building a data warehouse,

27:54

data models, product marketing analytics with sequel

27:56

on in these data models. I kind

27:59

of could get this started like as

28:01

a freelancer and send it as a

28:03

project on a project basis. Right. And

28:05

essentially I did that I started to

28:08

do this as a as my profession

28:10

I was selling as a freelancer and

28:12

the I was selling like a whole

28:15

package like I'm doing the platform. building

28:17

up like data breaks, big credits, whatever,

28:19

and data warehouse, interesting data, data modeling.

28:22

And I also saw the data analytics

28:24

piece, because I was doing as well,

28:26

like files and conversion rates and retention

28:29

rates calculation in industry, and it turned

28:31

out that the setup time of the

28:33

platform, it is I could get it

28:35

down to a couple of hours for

28:38

a startup that is starting out and

28:40

they are like greenfield. So I felt

28:42

like it's hard to justify for companies

28:45

to, you know, to ask for hours

28:47

of hours of my freelancing work for

28:49

this because it's not very fast. I

28:52

could do it on a project basis,

28:54

but then my project problem is a

28:56

marketing problem mostly. And I felt like

28:59

the analytics piece is essentially a product

29:01

problem because I can optimize ultimately automated

29:03

with the tool because, you know, as

29:05

I'm not sure information, but it's an

29:08

open source tool in the beginning. And

29:10

I already started to work on this.

29:12

And that was the time, like, actually,

29:15

what about if I just like make

29:17

this as a product and I stop

29:19

doing different freelancing, I just do this,

29:22

Mitsu as a product, and we are

29:24

going to charge on C, on C,

29:26

on C, basically how many seats we

29:28

are selling. So that was the idea.

29:31

And I take the deep breath and

29:33

just like jumped into it. I honestly,

29:35

I had, I would say it was

29:38

quite, it wasn't that hard of a

29:40

decision. Because financials were going to quite

29:42

good place after being a freelancer for

29:45

a couple of years. Plus I had

29:47

already invested friends. in my network that

29:49

were telling me like this is a

29:52

good story. Yes, you could already do

29:54

it. And like, you know, if you

29:56

get together a team, then you can

29:58

get funding and like, you know, like

30:01

a precede round, which is like the

30:03

earliest on the ground, you can get

30:05

started, just try it out, do it

30:08

for two years, it doesn't work out,

30:10

you can still go back to freelancing

30:12

and, you know, tell your hours or,

30:15

you know, what are, why not let's

30:17

started, let's started, and then I did

30:19

it, and then I did it, and

30:21

then I did it, and then I

30:24

did it, I did it, I did

30:26

it, I did it, and, and, and,

30:28

and, and, and, and, and, Yeah, it

30:31

was it's an action of two years

30:33

now that we are building the product.

30:35

Oh, wow. Very cool. Very cool. And

30:38

you're in the middle of a funding

30:40

round now if I heard you say

30:42

we are looking to get started. Yes.

30:45

Cool. Getting closer and closer. Yeah. Very

30:47

cool. Very cool. I think it's an

30:49

interesting story because I think there's a

30:51

lot of people that they're in that

30:54

case where they're either you're doing freelance

30:56

or they're doing like kind of this

30:58

thing and they see an opportunity. I

31:01

mean myself right I have a lot

31:03

of the back-end processes to post a

31:05

Franks world are through something I custom

31:08

wrote and kind of always in the

31:10

back of my mind I'm like this

31:12

would be a pretty cool sass right

31:15

like for content creators like so it's

31:17

kind of like one of those things

31:19

where but you know my my you

31:21

know I have the three kids and

31:24

I have all these things so like

31:26

it's always an excuse. I'll not do

31:28

something. So whenever someone is actually taking

31:31

the leap to jump forward and take

31:33

things to the next level, I always

31:35

want to know, like, what was the

31:38

thought process that went into that? Because

31:40

I know, one, a lot of people

31:42

are in that situation and they might

31:44

say, I am. And also, too, quite

31:47

frankly, selfishly, I want to know, like,

31:49

what is the, what's the mental math,

31:51

what's the mental model to execute on

31:54

that? That's cool. It's a valid excuse,

31:56

I would say. I have one and

31:58

it's it's tough enough. Right, right, right,

32:01

right. No, that's true. Yeah, yeah. Your

32:03

child is about the same age as

32:05

my youngest. and wow, this is a

32:08

wild age. My teenager, I love them

32:10

all, but teenagers are teenagers and he

32:12

has all sorts of opinions about what

32:14

we should be doing when the little

32:17

ones having a tantrum. And I was

32:19

like, I turned him and I said,

32:21

you know, like, when I was 15,

32:24

I had everything figured out too. But

32:26

here we are. So, you know, my

32:28

dad said something very similar to me.

32:31

when I was his age telling him

32:33

how I thought things should be done

32:35

and he kind of like yeah you

32:38

know there was a quote from an

32:40

old movie and I forget what it

32:42

was but it was this guy saying

32:44

that when he was 18 he thought

32:47

his father knew nothing and then by

32:49

the time he was like 20 to

32:51

23, he was amazed how much his

32:54

father had learned in the last four

32:56

or five years. So, like kind of

32:58

like, yeah, like now I'm going through

33:01

that and I'm like, oh, from the

33:03

other end, and it's like, I, I

33:05

feel like I've gained a whole new

33:07

level of understanding and of what my

33:10

folks went through. So, um, but by

33:12

the last thought about this, like, yeah,

33:14

I have a lot of things not

33:17

figured out. So it's like, so it's

33:19

like, yeah. to paint the full picture

33:21

is like, I took a lip of

33:24

faith with this, but I had to

33:26

take anyway with the pre-lancing. It was,

33:28

I would say it's almost equally hard

33:31

to decide on this, like, you know,

33:33

equal amount of stress to make a

33:35

decision. Building a company with employees and

33:37

everything is much harder, of course. I

33:40

would say it's a different set of

33:42

problems. But on the other hand, there

33:44

is if you have good talent in

33:47

your company, like we have, actually I'm

33:49

lucky to have my people, we are

33:51

six people that way. It's a good,

33:54

good, it gives me, like I sleep

33:56

better now than compared to when I

33:58

was a freelancer alone. you

34:01

know, if you are freelancing alone,

34:03

you have to do marketing, you

34:05

have to find your clients, it's

34:07

everything on you. Now it's like

34:09

kind of distributed, I do this,

34:11

my my co-founder is doing that,

34:13

you know, yeah, we forget to

34:15

mention, I think we mentioned the

34:17

beginning, I have a co-founder as

34:19

well, so I'm not alone in

34:21

this, you know, so we share

34:23

responsibility, I do this, if anybody

34:25

in the audience is thinking about

34:27

funding a company, I would say,

34:29

don't do it alone. Right, right.

34:31

Well, I think you also said

34:33

something very profound was like, you

34:35

don't have it all figured out.

34:37

And I think this, you know,

34:40

there is this myth that before

34:42

you do something, you have to

34:44

figure it all out. And you

34:46

know, all these self-improvement groups all

34:48

say more or less the same

34:50

thing. Like, you just have to

34:52

take the first step and you

34:54

don't want to see the whole

34:56

staircase. you know there's definitely a

34:58

something you said for that like

35:00

you know just get started you

35:02

know and a lot of things

35:04

like even this podcast was Andy

35:06

and I didn't know what we

35:08

were doing when we started and

35:10

you know here we are eight

35:12

years later you know you know

35:14

now we've got to figure it

35:16

out but and even then there's

35:19

still things we haven't figured out

35:21

right so one example is kind

35:23

of like how do we onboard

35:25

guests how do we onboard new

35:27

guests and you know like it

35:29

you know for you for you

35:31

the experience was smooth because that

35:33

we're still working on our automation

35:35

systems and things like that but

35:37

you're right it's a lot easier

35:39

too when you have someone else

35:41

because my other podcast someone does

35:43

all that type of aspect the

35:45

front facing stuff and and it's

35:47

definitely the logistics there I said

35:49

to convince her to do this

35:51

as well so yeah but so

35:53

Let's switch to the pre-made questions.

35:55

Hopefully you have them on the

35:58

screen because I think I think

36:00

I'd be very interested in hearing

36:02

your answers. Particularly like how'd you

36:04

find your way into data. Did

36:06

you find data or did data

36:08

find you? Oh, it's good. Actually,

36:10

my first job ever was related

36:12

to me. Like at that time,

36:14

probably the world's biggest amount of

36:16

data. I was I was working

36:18

for the large other collider in

36:20

ingenuity. Oh, no way. Very cool.

36:22

That was my first job. That's

36:24

over at CERN. Yes. Exactly. I

36:26

did four years there. I was

36:28

working on real-time analytics of physics,

36:30

which is like a funny thing

36:32

about this. You are doing it

36:34

only there, maybe in the US

36:36

and maybe a couple of places

36:39

like in China. Right. But that

36:41

profession is like you can sell

36:43

it on a free place in

36:45

the world. You're stuck with it

36:47

there. Right. Nevertheless, it was very

36:49

interesting dealing with data that is

36:51

essentially... handling light speed, speed of

36:53

light, you know, that things flew

36:55

around and the accelerator and you

36:57

have to, you have to monitor

36:59

that. All the protons are moving

37:01

in the accelerator, it's amazing. So

37:03

huge amount of data, amazing. I

37:05

was back in the engineer to

37:07

be, to be fair. And then

37:09

I joined one of the companies

37:11

in thermal industry and there I

37:13

was also a back and engineer

37:15

in the beginning, but like very

37:18

fast I shifted to data because

37:20

We were building a back-end service

37:22

for like the customers, you know,

37:24

for the front end essentially, and

37:26

we were dealing with millions of

37:28

visits and everything. So it's a,

37:30

and there was an opportunity to

37:32

me to become like an official

37:34

data engineer, and I felt like

37:36

this is like, this is like,

37:38

I need to do this. It's

37:40

much more interesting to do the

37:42

big data, they had data breaks

37:44

as well, so it was very

37:46

interesting to deal with that, that,

37:48

that, that, that, that, that, that,

37:50

that, that, that, that, that, that,

37:52

that, that, that, that, that, that,

37:54

that, that, that, that, that, that,

37:56

that, that, that, that, that, that,

37:59

that, that, that, that, that, that,

38:01

that, that, that, that, that, that,

38:03

that, that, that, that, that, that,

38:05

that, that, that, that, that, that,

38:07

that, that, to do it to

38:09

faith again and become a data

38:11

engineer. Turned out to be I

38:13

was the only one for the

38:15

whole department. at that time yeah

38:17

yeah it was so was your

38:19

was your background in physics or

38:21

was your background in computers tech

38:23

in software okay actually actually oh

38:25

really yeah but it's no nothing

38:27

compared to the ones that you

38:29

have now you know you didn't

38:31

like there was nothing like it

38:33

was 2010 I finished Oh yeah,

38:35

AI was a different industry than

38:38

they probably called you a data

38:40

scientist was probably were business intelligence

38:42

or something like that. Like it

38:44

was, it had a different names

38:46

then. It's been, that's the thing

38:48

with predictive AI, it's been around

38:50

for decades, it just changed names

38:52

and, but yeah, AI is a,

38:54

is an interesting, it's a lot

38:56

more crowded now than it was

38:58

in 2010. Oh yeah, it's different.

39:00

Back then it was classification that,

39:02

that's a big deal, big deal.

39:04

And, actually, it was 2012, like,

39:06

like, like, like, like, like, like,

39:08

like, like, like, like, like, like,

39:10

like, like, like, like, like, like,

39:12

like, like, like, like, like, like,

39:14

like, like, like, like, like, like,

39:16

like, like, like, like, like, like,

39:19

like, like, like, like, like, like,

39:21

like, like, like, like, like, like,

39:23

like, like, like, like, like you

39:25

know we and I immediately like

39:27

abandoned it because I felt it's

39:29

like this is like you know

39:31

it's a research thing there is

39:33

no really business on the of

39:35

course there was but not in

39:37

Europe and I became a better

39:39

kind of engineer essentially and then

39:41

I fell in love with data

39:43

I fell in love the database

39:45

actually I saw I love that

39:47

platform so much I'm doing a

39:49

bit of advertising for them but

39:51

I think it's amazing I'm a

39:53

big fan too. I love snowflake

39:55

as well. Honestly, it's great. Everything

39:58

is in data because I like

40:00

that you are full in full

40:02

control of everything to the like

40:04

to the lowest level essentially the

40:06

files and it's open source base

40:08

spark. You know, it's I love

40:10

that it's not like I like

40:12

it's it's simple and everything and

40:14

you know, you can you can

40:16

get started with like a fun

40:18

quick and you are on in

40:20

data because it's a bit get

40:22

it's harder to get started, but

40:24

it's like you have the full

40:26

control. So anyway, I like it.

40:28

Bigfry is great as well. It's

40:30

improving a lot in recent years.

40:32

I'm promoting a lot, actually, among

40:34

our customers. So I think they

40:37

are doing a great job as

40:39

well at Google. Very cool. What's

40:41

your favorite part of your current

40:43

job? Oh, good question. Really enough,

40:45

sales? Interesting. Yeah, so I'm not

40:47

doing development anymore. I'm managing developers

40:49

because we are a small team

40:51

and I need to do that.

40:53

I'm doing the product vision, which

40:55

is I like it still, but.

40:57

The the best part is what

40:59

I like most is sales when

41:01

it works obviously I mean it

41:03

doesn't work it's I know sales

41:05

is sales gets a bad sales

41:07

gets a bad from non sales

41:09

people but it actually can be

41:11

quite fun like and obviously there's

41:13

also very bad days but there

41:15

are also very very good days

41:18

yeah and yeah and yeah but

41:20

there are also very very good

41:22

days yeah and yeah and yeah

41:24

the bad days. You know the

41:26

biggest mine shift I had and

41:28

I think it was the best

41:30

thing that so far I think

41:32

it came out from me shifting

41:34

to become like a founder is

41:36

that when you are like individual

41:38

contributor like a backend or data

41:40

engineer and you are working alone

41:42

in a bubble and in your

41:44

little space and you know you

41:46

can focus on one thing everything

41:48

the feedback loop is like you

41:50

know instant seconds and everything and

41:52

when I started to do sales

41:54

for the business, the feedback loop

41:57

is like sometimes six months long.

41:59

So like you write to a

42:01

company, they're ghosting you, you know,

42:03

and half a year later they

42:05

show up in your radar, they

42:07

sign into your application, self-service, not

42:09

book the meeting, like you know,

42:11

we have on the website, and

42:13

then they contact you like, I

42:15

remember these contacted us, it wasn't

42:17

relevant back then, but now it

42:19

is relevant, you know, right. We

42:21

just had this issue of what

42:23

you mentioned with a huge amount

42:25

of data with the current incumbent

42:27

solutions now the subscription We should

42:29

renew the subscription but we look

42:31

at the bill they are proposing

42:33

us and it's impossible to pay.

42:35

We also have this data warehouse,

42:38

you know, let's let's collect and

42:40

let's see if it works for

42:42

us, you know. And it's such

42:44

a good feeling to have, you

42:46

know, I mean, we are six

42:48

people and companies of like a

42:50

thousand employees come to us with

42:52

this, this message. And it's such

42:54

a, I love this feeling when

42:56

you have this like, you know,

42:58

you are a bit validated and.

43:00

No, that's a great way to

43:02

put it because there's definitely like

43:04

this aspect of, you know, it's

43:06

such that with coding and kind

43:08

of regular things that the feedback

43:10

loop is far tighter. Like I

43:12

noticed that my hobbies changed. Like

43:14

I picked up archery during the

43:17

pandemic because, you know, the feedback

43:19

loop on that is instantaneous, right?

43:21

As soon as you let go,

43:23

like. you know you either hit

43:25

it or you didn't you can

43:27

kind of also mentally rehearse like

43:29

well did I let go wrong

43:31

did I aim wrong did I

43:33

breathe right like like it's instantaneous

43:35

whereas like and like I wasn't

43:37

always into that when I was

43:39

you know prior to joining sales

43:41

where the sales cycles are longer

43:43

and things like that so I

43:45

think I think you're right like

43:47

it's it's but but you also

43:49

get that good feeling when they

43:51

come back to you later and

43:53

you you solve their problem in

43:55

sales is ultimately There's a guy

43:58

called Dean Graziosi. He's one of

44:00

those self-help gurus that I kind

44:02

of subscribe to and he's like

44:04

sales is really service right like

44:06

if you start looking at that

44:08

way and You know you have

44:10

all these pictures in your mind

44:12

you know like like the movie

44:14

from Grand Gary Celine Ross of

44:16

how it is and but then

44:18

you think about it it's really

44:20

about service you're helping them solve

44:22

a problem and you have to

44:24

guide them to realize like you

44:26

are the solution to their biggest

44:28

problem and when you when they

44:30

come around to it it's a

44:32

very satisfying feeling. Oh yeah definitely

44:34

the last two years basically it's

44:37

learning that for me obviously there

44:39

is a lot of other things

44:41

like hiring and product, roadmap building

44:43

and all these managing investors because

44:45

you have to manage them. as

44:47

well. It's all that, but the

44:49

most fascinating is that one that

44:51

I had, and the thing I

44:53

like in it is that it

44:55

changed me in the person that

44:57

I'm accepting this half a year

44:59

feedback loop and I'm contracting myself

45:01

in that that, okay, they didn't

45:03

immediately buy the solution, but we

45:05

have a logo now that we

45:07

can come back to, you know,

45:09

that's one more account on our

45:11

like CRM, you know, we can

45:13

come back to and that eventually

45:16

they might turn in and what

45:18

I see now after two years

45:20

that. the pipeline, you know, currently

45:22

is filling up with logos or

45:24

accounts that were contacted last year,

45:26

you know, like last January, more

45:28

than one year ago, and they're

45:30

coming back, and it feels like

45:32

if I do this for the

45:34

next three, four years, my time

45:36

will be filled with meetings. Right.

45:38

Yeah, and it's like, okay, this

45:40

is actually maybe, maybe it's not

45:42

work, you know. Right. This product

45:44

is maybe going to be sold

45:46

like to some, to the big

45:48

clients and everything. It's the whole

45:50

thing is feasible as a product,

45:52

you know, or as a business.

45:54

I just have to take my

45:57

time and build this pipeline, build

45:59

this sales, sales machine essentially. In

46:01

contrast, when I started out, it

46:03

was, oh my God, somebody like

46:05

replied to me, but they didn't

46:07

show up on a meeting. Oh

46:09

my God, this is end of

46:11

the world, you know, I'm doing

46:13

something wrong, I should stop this.

46:15

And I'm like totally fine with

46:17

it now and you know I

46:19

just got used to it and

46:21

it's yeah you have the confidence

46:23

now in this whole thing again.

46:25

With that said we don't have

46:27

everything figured out. Right right right

46:29

I think that's important to realize

46:31

like you don't have to have

46:33

it all figured out right? Yep.

46:36

Thankfully we only have to live

46:38

one day at a time and

46:40

one moment at a time because

46:42

otherwise it could be very overwhelming.

46:44

We have a couple of complete

46:46

sentences. So three, actually. So the

46:48

first one is, what I'm not

46:50

working, I enjoy. Yeah, being with

46:52

my family, kids, cool. My kid,

46:54

my kid. We play a lot

46:56

of Lego. No, that's cool. That's

46:58

cool. Yeah, it's a fun age.

47:00

It really is a funny. It's

47:02

a challenging age. Yeah. But it's

47:04

fun. I look forward. The next

47:06

one is, I think the coolest

47:08

thing in technology today is. I

47:10

would say in the data world.

47:12

Of course, AI. will dodge this,

47:14

I train if you don't mind.

47:17

Yeah, everybody knows that. For me,

47:19

the most fascinating thing is data

47:21

lakes, definitely. Yes. You know, that

47:23

it's a paradigm shift in processing

47:25

and the whole how things are

47:27

done, you know, storage is virtually

47:29

free, computation, doesn't have to be

47:31

running all the time. Right. You

47:33

can make it, you can make

47:35

it super cost efficient, super, actually

47:37

in a company with it. You're

47:39

trying to add a huge amount

47:41

of data. Any amount of data.

47:43

Yeah. So that seems, for me,

47:45

the data engine, this is like,

47:47

it's a paradigm shift essentially. Interesting.

47:49

Yeah. And I think the big

47:51

paradigm shift has been like the

47:53

schema for decades was the thing

47:56

you started with. And it was

47:58

the Andy, Andy. My co-host has

48:00

a really good way of putting

48:02

it. Shout out to you Andy,

48:04

because I know you're listening later,

48:06

is the idea that the primacy

48:08

of the schema, right, where the

48:10

schema kind of runs everything. But

48:12

now with data warehouse, data, data

48:14

lakes, like it's really about the

48:16

data, right? And the schema comes

48:18

later. And the cool thing about

48:20

that is you can capture stuff

48:22

today. that you may not think

48:24

is important but somewhere down the

48:26

road it turns out that that

48:28

was important some extra factor that

48:30

you didn't really think about that's

48:32

a that's a good answer because

48:35

everybody says AI this AI that

48:37

self-driving car yes of course, of

48:39

course, but you know, I think

48:41

I wouldn't say anything new with

48:43

that. And if I would exclude

48:45

that from the equation, the second

48:47

best thing to develop, right, is

48:49

sometimes like, it's like, right, and

48:51

the third one is data leaks,

48:53

you know, and, yeah, well, it's

48:55

true, like, it's funny, like, there's,

48:57

there's a term for that where

48:59

you can kind of remove all

49:01

the noise, I forget what it's,

49:03

I forget what it's, spoken language

49:05

in any given state, right? Most

49:07

states is going to be Spanish,

49:09

right? But the third most language

49:11

spoken tells an interesting story too,

49:13

right? So like in Louisiana, it's

49:16

French, in, you know, New Hampshire,

49:18

Vermont, it's French, right? Like, I

49:20

think in Texas, it's actually German,

49:22

which I think was interesting. Like,

49:24

so there's a lot of interesting,

49:26

and there's a whole historical reason,

49:28

Texas, Germans, and things like that.

49:30

But like, it's very fasting to

49:32

see. because there's you know obviously

49:34

the the elephant in the room

49:36

here in our industry is AI

49:38

but like what's the second you

49:40

know thing there tells another story

49:42

and I look forward to the

49:44

day when I can use technology

49:46

to blank that's the third one.

49:48

Oh my god this is a

49:50

good one again the high train

49:52

would be replaced my day to

49:55

day work. You know, but that's

49:57

that's a good one. I haven't

49:59

thought about this like what maybe

50:01

I will translate this question what

50:03

annoys me in my work and

50:05

what I would like to be

50:07

replaced, you know, I am really

50:09

bad at hiring. I'm okay to

50:11

say yes or no in the

50:13

end, but doing the process. I

50:15

hope. My co-founder and AI as

50:17

well will take over my hiring

50:19

tasks for the company. That's for

50:21

sure. Or I need to get

50:23

better in it very fast because

50:25

yeah, it's something I always pushing

50:27

away for myself. Although I shouldn't.

50:29

Interesting. Yeah, dealing with people I

50:31

think is always going to be

50:33

the interesting part of the equation.

50:36

So share something different about yourself.

50:38

Remember it's a family podcast? So

50:40

it's like, what's a question? So

50:42

it's like, what's a different, what's

50:44

a, what's a, what's a unique

50:46

hobby that you have that people

50:48

wouldn't necessarily think of or? You

50:50

know, with the kid, it's a

50:52

bit hard, but I used to

50:54

do work climbing. Oh, really? Oh,

50:56

very cool. Yeah, when I was

50:58

younger and I used to live

51:00

in France. near the Alps and

51:02

we did. Oh yeah, when you

51:04

were discerned presumably. Exactly. Yeah, in

51:06

certain I worked in Germany, Geneva

51:08

in Switzerland, but we were living

51:10

actually on the other side of

51:12

the border in France, it was

51:15

much cheaper obviously, so it's a

51:17

great thing to do. And we

51:19

did like her climbing every week

51:21

multiple times, it was amazing. I

51:23

love that. I like cycling a

51:25

lot, actually. I cycle in the

51:27

city. And Budapest is a very

51:29

good place to be cycling. It's

51:31

the best native community here with

51:33

the bicycle. hiking as well there

51:35

is some hills around you you

51:37

can do out and you know

51:39

do some do some hiking and

51:41

you know just going around walking

51:43

with the kid but that's you

51:45

know like nothing crazy I'm not

51:47

doing like high surfing or something

51:49

like that because right right right

51:51

right right that's cool that's cool

51:53

it's always fun to learn about

51:56

people's hobbies like for me it's

51:58

archery or you know you know

52:00

yeah so that's cool where can

52:02

people learn more about Me too.

52:04

They can come to our website.

52:06

there is most of the information

52:08

we have documentation page as well

52:10

if people want to look at

52:12

like how things are you know

52:14

how we are integrating to death

52:16

warehouses they can just contact me

52:18

okay I can find my my

52:20

there is a email form in

52:22

the in the applicant in the

52:24

website as well okay there is

52:26

like a book a meeting button

52:28

I'm happy to chat with people

52:30

who are interested oh very cool

52:32

yeah yeah So yeah, I'm happy

52:35

to happy to tell about this.

52:37

We have a YouTube channel that's

52:39

going to be renewed soon. Oh,

52:41

very cool. Yeah, we found a

52:43

very good AI tool that it's

52:45

like, you know, this is a

52:47

great, great use case for AI

52:49

to do explain the videos and

52:51

like onboarding videos. Oh yeah, you'll

52:53

tell me what tool that is,

52:55

because I can totally use that.

52:57

It's called Video scriptter. Okay. It's

52:59

a it's the primary use case

53:01

is documentation on board videos. Video

53:03

voice. You know, for SAS tools,

53:05

it's a perfect find for us.

53:07

You know, and by the way,

53:09

it's a company from Hungary as

53:11

well, from Budapest. Oh, very cool.

53:14

Yeah, a friend of us is

53:16

doing it. I'm not trying to

53:18

promote it. It's just, no, no,

53:20

no, no, like, I, yeah. I

53:22

think that the impact of AI

53:24

on video, I think can't be

53:26

understated, like, like, I use opus,

53:28

opus, opus, opus, opus, and I'm

53:30

actually an affiliate now so if

53:32

you want to support the show

53:34

click on that link folks but

53:36

it's amazing like it basically takes

53:38

long form videos like this like

53:40

this is going to be put

53:42

through opus and it'll chop it

53:44

up into like little YouTube shorts

53:46

oh and it's amazing like the

53:48

the the amount of views you

53:50

get on on that it also

53:52

exports to TikTok I'm not on

53:55

TikTok But like it's it's it's

53:57

impressive what it produces like and

53:59

I did so well like I.

54:01

the amount I'll I'll share you

54:03

the analytics after you know in

54:05

the in the post of recording

54:07

virtual green room but it's been

54:09

phenomenal I've had it for a

54:11

month I've gotten another thousand subscribers

54:13

like it has been ridiculous ridiculously

54:15

well yeah so I'll mention that

54:17

to you and then finally audible

54:19

is also speaking of sponsors and

54:21

supporting the show audible as a

54:23

sponsor I don't know if you

54:25

do audio books but any book

54:27

that you recommend I have

54:30

audible actually. Okay, cool. I listened

54:32

mostly to these like startup books.

54:34

I think the latest one. I

54:36

listened mostly to these like start

54:38

up, you know, books. I think

54:40

the most the latest one was

54:42

founded at sales, I think or

54:44

something about that. Okay. Yeah, it's.

54:46

Actually, maybe the title is not

54:49

this, but it was something about

54:51

around that one. I used to

54:53

listen to a lot of audio

54:55

books from from Aldewal. Yeah. Is

54:57

it CEO led sales? I have

54:59

to look it up. Sorry. No,

55:01

that's fine. That looks good too,

55:03

actually. Now that I see it.

55:05

There is. I also listen to

55:07

a lot of business books and

55:10

kind of motivation stuff. Exactly. You

55:12

know, it's great when you have

55:14

a little one because you can

55:16

kind of like earbuds in and

55:18

kind of like. you know, when

55:20

you're not paying directly attention to

55:22

them, but you want them to

55:24

keep them alive, it's great for

55:26

that. Especially with boys, like you

55:28

spend a lot of time just

55:30

keeping them alive. I don't know,

55:33

I will say. And I know

55:35

people are going to be like,

55:37

it's true for girls too, but

55:39

I'm sure it is, but I

55:41

have three boys, so that's all

55:43

I know. And finally, any parting

55:45

thoughts. Thank you so much for

55:47

the opportunity to speak. I really

55:49

enjoy this. Yeah, me too. My

55:51

parking thought is like, the more

55:54

I speak about this, it's the

55:56

better it gets, actually. Yeah. Well,

55:58

that's just it. Yeah, no problem.

56:00

I mean, that's just it. I'm

56:02

trying to encourage more people to

56:04

do podcasting, particularly on my team

56:06

at work. But they're like, well,

56:08

you're so smooth, you're so good

56:10

at this. And I'm like, well,

56:12

thank you. awful, but like it

56:15

just, you just have to do

56:17

it, right? It's like anything else,

56:19

right? Like, you know, very few

56:21

people are naturally born gifted in

56:23

anything. I mean, Michael Jordan, right,

56:25

globally known as one of the

56:27

best basketball players of all time,

56:29

if not the best. He was

56:31

kicked off his team when he

56:33

was in middle school. Oh, wow.

56:36

He didn't make the team, right?

56:38

Yeah, and he was, you know,

56:40

so he basically... practiced to the

56:42

point where, you know, he became

56:44

Michael Jordan. We know today and

56:46

everybody in the world knows who

56:48

he is, right? But like very

56:50

people realize like when he was

56:52

like middle school or whatever, like

56:54

he didn't make the team. So

56:56

it's just kind of like this

56:59

goes to show you like you

57:01

know, I'm not saying everybody's going

57:03

to be Michael Jordan, but you

57:05

can be the Michael Jordan in

57:07

your particular industry or whatever. So

57:09

with that inspirational quote, we'll let

57:11

Bailey finish the show. And that's

57:13

a wrap on another data-fueled episode

57:15

of data-driven. Big thanks to Istven

57:17

Mazoros for sharing his journey from

57:20

CERN to SAS and for proving

57:22

that even analytics tools can have

57:24

personality and the occasional feline alter

57:26

ego. Before you go back to

57:28

fine-tuning your SQL or debating data

57:30

lake architecture on slack, don't forget

57:32

Impact Quantum is back. Yes. The

57:34

podcast that explores the quantum frontier

57:36

with just the right amount of

57:38

existential dread and nerdy wonder has

57:41

officially relaunched. Search for Impact Quantum

57:43

on your favorite podcast app or

57:45

Visit Impact. quantum.com to

57:47

subscribe. And hey if you

57:49

if you enjoyed

57:51

today's episode, or

57:53

you just like

57:55

the sound of

57:57

my synthetic British

57:59

accent, please rate

58:02

and review the

58:04

show. show. It It

58:06

helps more curious

58:08

minds find us us

58:10

keeps the data

58:12

vibes flowing. flowing.