sds podcast episode 123 with rico meinl · kirill: this is episode number 123 with the unstoppable...
TRANSCRIPT
Kirill: This is episode number 123 with the unstoppable Rico
Meinl.
(background music plays)
Welcome to the SuperDataScience podcast. My name is Kirill
Eremenko, data science coach and lifestyle entrepreneur.
And each week we bring you inspiring people and ideas to
help you build your successful career in data science.
Thanks for being here today and now let’s make the complex
simple.
(background music plays)
Welcome back to the SuperDataScience podcast. Today I
have one of the most inspiring episodes that you have ever
heard on this show. We've got Rico Meinl calling in from
Germany and this person is a machine. He's got passion,
he's got drive, he's got that whole concept of reckless
commitment down pat. So Ben Taylor talks about reckless
commitment, and that is the notion of just committing
yourself to something that you feel is impossible that you
will do, and you still commit to it, and you go, and once
you're committed, you go and do it. So just an example, Rico
has just in a month set up an AI meetup group and hosted a
meetup event for 45 people in Hamburg, Germany. He went
to his executives in his company and suggested that they set
up an AI department in order to augment their operations
with artificial intelligence, and now he's doing that. He's also
studying at the same time. He's also gotten himself a mentor
who is guiding him through his career.
So a crazy amount of incredible things that he's done in a
short period of time, and this is a person that you can, and
anybody can, learn a lot from just through thinking about
his attitude towards data science, AI, and just anything he
does in life. Also important to note that this podcast is
available in video version, so if you go to
www.superdatascience.com/123, you will see our whole
conversation in video. So if you have the opportunity to
watch this on your laptop, if you're at home right now and
you can just switch your laptop or your phone on, then go
and do that. The podcast is on YouTube, and you'll see our
whole conversation there, you'll see us laughing. But if not,
if you're just running, or in the car, or anything like that,
then just keep going with this audio, this is definitely worth
it. You will get tons and tons of value.
So all in all, it's going to be an incredible ride, so brace
yourselves and off we go. I bring to you the unstoppable Rico
Meinl.
(background music plays)
Welcome ladies and gentlemen to the SuperDataScience
podcast, today I've got a super exciting guest on the show
calling in from Germany, Rico Meinl. Rico, welcome to the
show. How are you going?
Rico: Thank you, Kirill. I'm doing pretty good.
Kirill: That's awesome, and great to see you again. We initially met
back at DataScience GO in October, it's been quite some
time, and you've accomplished some crazy things in those
months. I'm really excited to talk about it. But first, I wanted
to start off to get into this space and as an ice breaker, what
were we talking about just now, about being nervous versus
being excited. What did you think about that whole concept?
Rico: I think it's interesting because it's sort of like what you
mentioned with the TED Talk, it's always the same feeling,
and I think that's really true, because sometimes when I'm
nervous before a presentation or an important talk, once I
realise that it's actually just something that I really care
about and that I'm excited to do, so there was the same
feeling and it's not really something that holds you back if
you feel like doing so.
Kirill: Yeah, and we were just talking about being nervous, and I
actually watched a TED Talk where when you're nervous
and when you're excited, you experience exactly the same
feelings, where you have some specific type of breathing, you
sweat a little bit, you're anticipating what's going to happen.
That happens when you're nervous, if you think about it,
and that happens when you're excited. So if you train your
body – this was a TED Talk by Simon Sinek, (oh, this was
one of his videos, but not his TED Talk) and he was actually
giving an example of athletes at the Olympics. The reporters
always ask them, "Were you nervous, were you nervous, are
you nervous?" and they never say they're nervous. They
always they're excited. Because they've trained their bodies
to be excited. But you also mentioned a cool thing. What
Kyle C said, right? Can you repeat that? About the technique
that he gave you that you're using now for yourself.
Rico: Oh yeah, it's just what he recommended at his talk was that
whenever you're really excited or nervous about something,
you have this voice in your head that keeps telling you the
things you might do wrong and where you might fail. So
what I do now before important presentations or something,
I try to talk to the voice, like he recommended. Talk
everything down and answer the voice in your head so you
really calm it down. And then what's also interesting in that
sense is what you mentioned at the DataScience GO
conference, what you're going to ask yourself is, what is the
next thing I'm going to think about? And then your mind is
empty. You don't think about anything at that moment. And
that really helps me sometimes to calm myself down.
Kirill: Nice, nice. Very cool. So are you nervous, or are you excited
right now?
Rico: I'm excited!
Kirill: Awesome. Ok, well for the benefit of our listeners, I'm going
to actually just recite in my own version the email that you
sent me. So I met Rico at DataScience GO, it was October,
and now it's been what, 2 months, literally since then. No,
that was November, what am I talking about? It was
November. So it's been like a month since DataScience GO,
and Rico is crazy. So first of all, Rico flew all the way from
Germany to San Diego just for this conference. Is that right?
Or am I getting this wrong?
Rico: That's right.
Kirill: Man, that is crazy. A huge thank you for that. There were a
couple of people who came from all over the world, so I was
very inspired by that. And then a month later, after the
conference, I get this email from Rico, and he's like, "By the
way, Kirill, I wanted to say—" Oh, no, hold on. At the
conference, what did you ask me at the conference?
Rico: I was actually inspired, and it was a lot about setting goals,
right? So what I realised during the talks, that these people
are doing amazing things, and are really inspiring. So what I
wanted to set myself as a goal is I want to reach so far that
next year, I can be one of the presenters at the DataScience
GO conference, so the one in 2018. So I approached you and
I asked, "Kirill, I really want to be a presenter. And then
what you said was, “Well, first you have to succeed.” That
was a perfect response for my type of personality, I think. So
that really motivated me.
Kirill: Nice. Very nice. And I remember, I haven’t told you this, but
I remember I was sitting at the back of all of the chairs, at
the back there’s these seats—because I remember at the
moment where we had all the audiovisuals, the recording
team, the camera crew and so on. And I was sitting behind
on my laptop, I was talking to someone, and then Rico rocks
up. First of all, you’re very tall and I’m like, “This guy, what
does he want from me?” This is towards the end of the
evening, I think first day or second day, I was already a bit
tired. I was like, “What does this guy want from me? I’m
tired.” And you’re like, “I want to be a presenter.” I’m like,
“Dude. You bet.” And I’m like, okay, that’s cool, I’ve got to
get myself together and give him an appropriate response.
When would I want someone to present at this conference?
When he succeeds, when he has a successful story to share.
“You’ve got to succeed first.”
And that’s the end of our—of course, we chatted more, but
we didn’t talk more about this. And so then I get back, a
month passes, and I get this e-mail, and Rico is like, “So I
got back to Germany. By the way, Kirill, do you remember
we had this chat about me being a presenter? I’m not e-
mailing you about that at all. I hope you remember it, but
basically, I got back to Germany, I set up a Meetup group for
artificial intelligence which has like 250 people in it already,
I’ve recommended my company to incorporate artificial
intelligence in their operations and now we’re setting up an
AI department. And also I reached out to one of the speakers
at DataScience GO, Richard Hopkins, who is my mentor.”
You basically stole my mentor and now Richard Hopkins is
your mentor and you’re catching up with him on Facetime. I
mean, you caught up with him twice in a month. I don’t see
him that many times in a quarter. So I’m like, “Whoa, this
guy is on a role,” and I had to invite you here and learn more
about this. And the other thing is you got inspired by some
of the speakers and now you’ve shifted your sleeping habits
to accommodate all these crazy things you’re doing. So, very,
very inspiring accomplishments. I don’t even know where to
get started? Where are we going to get started with this,
Rico?
Rico: Let’s start at the very beginning.
Kirill: Okay. Let’s start at the beginning. Walk us through it.
Rico: All right. Maybe a little background.
Kirill: Yeah.
Rico: Okay.
Kirill: First of all, how old are you? If you don’t mind sharing, how
old are you?
Rico: I am 20 years old.
Kirill: Okay. That’s very impressive. Go ahead.
Rico: So, when I was 18 I finished high school and I wasn’t really
sure what I wanted to do, so I went to Canada for a year
because I always wanted to travel the U.S. So that’s what I
did, I started working at a restaurant in Toronto as a
dishwasher actually. It was a really exciting time. I also went
to Vancouver for half a year to work in a ski resort, travelled
America on the way, and then travelled America again, and
finished off with two months and then went home to
Germany.
And over the course of that year, I really thought that I was
passionate about movies. I really thought I wanted to make
movies. I found this little university in Hamburg that has a
study called Media and Computer Science. And there was an
article online which stated that some people that studied
here, they eventually went to the film industry to do movie
effects and graphics for movies.
Kirill: Okay. I was thinking you wanted to become an actor or
something like that.
Rico: No, no. (Laughs) More on the technical side. So that’s what I
actually wanted to do. I came here, I started studying, and
I’m doing like a co-op program. Is that a thing in Australia
as well?
Kirill: Explain it a bit more. A co-op is not university, it’s before
university – is that what you mean?
Rico: No, it’s more like I’m enrolled as a student and then in my
semester holiday, because in Germany we have semesters of
six months, so I study for 3 months and have exams and
then everyone has 3 months of semester holidays. So, in
that time I am working at a company and I have a contract.
And also I will stay at the company for two years after my
study. So I think it’s a really great deal that goes both ways.
Actually, at my initial interview, my now-COO asked me,
“You are aware of the fact that this study subject is 80%
computer science?” And I was not. (Laughs) So I was like,
“Yeah, sure.” So he asked me, “Do you think you can do it?”
and I was like, “Oh, yeah, absolutely.” So I started studying,
just diving into coding, because there’s a lot of practical
applications at the school. In the second semester, after my
first practical experience in the company, I decided, “I don’t
want to do movies anymore, I want to do computer science.
There’s way more opportunities.”
So, yeah, I’ve been doing that and there was just a general
curiosity about AI. So for my second practical semester in
my company, I asked to go into the customer service
software department, where we have a chatbot, which is
some kind of AI in that sense because it’s a rule-based
model. Yeah, so that was basically not an excuse, but some
sort of excuse to get involved with AI.
So, I went to Udemy because we have a company’s account
where we can do courses for free. So I was like, “Are there
any courses on AI that I can do?” And I was surprised
because there was, and it was your courses. Then over the
course of my practical semester, I did the machine learning
and deep learning and AI course, and some other ones as
well.
Kirill: Which ones did you like the most?
Rico: I liked the deep learning courses the most.
Kirill: Yeah? Did you like how it starts off where they’re saying
what is the Internet, the 1994 video clip and they didn’t even
know what the Internet was on one of those CNBC type of
shows? I like that course because of the style we put into it.
Sorry for the side note. Okay, deep learning, I agree. I like
that course a lot.
Rico: Yeah, so then I was actually in Edinburgh on some sort of
vacation and I saw that you guys were doing the
DataScience GO conference. I was like, “Why not? I’m just
going to do it.” So I bought the ticket—
Kirill: When did you buy the ticket?
Rico: In August.
Kirill: In August? So like a few months before the conference?
Okay.
Rico: Exactly.
Kirill: And then you’re like, “Screw it. I’m just going to fly half the
way across the world just to come to a conference in San
Diego.”
Rico: I had this idea that I thought that the whole thing is going to
be really expensive, the ticket and the flight, but I thought,
“If I’m going to meet one person that inspires me, if I’m going
to meet one person that is going to make an impact on my
life, it’s going to be worth the investment.”
Kirill: Oh, nice. I love that thinking. Did you meet at least one
person?
Rico: I met more than 20 persons.
Kirill: (Laughs) Nice!
Rico: I met so many amazing people at that conference.
Kirill: Yeah. That’s so cool. And one of them is your mentor now.
That’s crazy! I’ve got to chat to Richard about this. I’m going
to send him this video. Richard, if you’re watching this, hey
man, this is fate. Okay, cool. So then you got there, you got
inspired, but what often happens is people get inspired, but
they don’t do anything about it. It’s cool, they have this
feeling, and then—you get most inspired at an event, you
know. You get inspired after reading a book and so on, but
at an event, you are surrounded by these people for several
days, you are committing time and money into travel, but
then it fades off inevitably within a week, or maybe for some
people within a few months. How did you make yourself
actually follow through on your plan?
Rico: Yeah, what you mentioned is definitely true. I’ve been there.
What I realized, which works for my type of personality, for
example, if I want to do something, like a course, I buy it
first. And then I sort of have to do it eventually. That was the
same thing with my presentation in a company and also the
Meetup. So when I had this idea at my company, which
we’re going to get to later, I guess, I thought I was going to
do it. So, the first thing I did before having anything to
present, I texted my business unit manager and I was like, “I
really want to present this thing. Are you going to be free
next week or something to have a meeting?” And she was
like, “Sure,” so then I created a presentation.
Same thing with Meetup. I came home from DataScience GO
and the same night I created a Meetup. So once the site was
up there and people started signing up for it, I was like,
“Now I have to do it.” So we had our first meetup last
Saturday.
Kirill: Nice. How many people attended?
Rico: It was 45 people attending, so less than expected, but I
guess that’s what you learn from those kind of events – it
doesn’t always go as you hope it would go, because 115
people signed up for it. But it was so great, it was certainly
an experience for me. I’ve never talked in front of so many
people and I really liked it.
Kirill: That’s so cool. You went from attending a conference to
hosting your own event within a few weeks. That’s crazy,
man. And you said the same thing with Richard. How did
you manage to make that happen, because Richard is not an
easy guy to convince to be your mentor?
Rico: I got involved in a talk with him on this Saturday night at
the networking. It was just really interesting talk. To be
honest, I don’t remember exactly what we talked about, just
about where I came from and where he came from, and then
he told me that he was really inspired by my story and then
the next morning he told me that also, because of the whole
chain of cause of the conference, he completely redesigned
his talk, right? So he ended up talking a lot about
mentorship.
At one point I stood there and I realized I had a really great
connection with him the night before, and then he was
talking about how important it is to get a mentor and all
that. So really after his talk, I went out there and I was like,
“Richard, I really want you to be my mentor. Do you think
we can work anything out to have like a Skype mentorship,
because I live in Germany?” I’d love to visit Australia one
day, because I’ve never been. So he said absolutely, we kept
in touch, and we had our first talk I think a week or two
after the DSGO.
Kirill: That’s really cool, man. Richard is a great mentor. For those
who don’t know, Richard Hopkins used to be a Director at
PwC, or even higher than a Director, for structuring and
business turnaround, while now he is a CEO at a huge
lettuce growing company located in Tasmania. So, yeah,
man, that’s really cool.
And how are these catch-ups going? Is it hard? Because
Richard is the type of guy—he’s not going to follow you, you
have to take control. When he was my mentor, that’s the
first thing he told me: “Kirill, I’m happy to be your mentor,
but you have to be the one driving this thing. I have lots of
things going on, and if you can’t be bothered to set up a
meeting with me, then it’s not my responsibility.” So how
has that been? You said you’ve had two catch-ups with him
so far. How do you go about it? Especially a remote mentor.
When he was my mentor, we were in one city, we were
chatting all the time, we’d go out for lunch. How does it work
for you guys?
Rico: We set up this first meeting over e-mail and I think
LinkedIn, and then what I really liked, for the first meeting
he took over the control, which was great for me because I’ve
never been mentored before. So he took it over and he said,
“Listen, Rico, this is how I usually approach these things.”
So it was a good introduction for me on how things are going
to go, we had a really interesting talk, and then we had our
second meeting, and that’s when I realized what you’re
saying, that I have to make sure that I have questions
prepared and ask for stuff.
Because of course, I actually greatly appreciate that he takes
his time even though he is busy. I think it’s also great that
we’re not wasting time, so we catch-up, talk about some
stuff, I ask my question and if it’s only 20 or 30 minutes,
only that. Of course, because of the change in my job now,
I’m trying to build this AI research lab up, I had a lot of
questions regarding that, and he was able to help me a lot
with that.
Kirill: Nice. Okay, during Richard’s talk, he emphasized this one
important point that mentorship has to be a two-way street,
that you cannot just take, take, take from your mentor. Your
mentor has to get some value out of it as well, at least even if
you get 90%, he gets 10%, or 80%-20% or something like
that. Because it’s great to help other people, but you also
want to grow yourself, you want to learn. So, my question to
you is, what are you contributing to this relationship? Just
out of curiosity.
Rico: I actually asked him the same question because I was
interested. Before we first talked, I wasn’t actually sure what
I was going to contribute. But he said he’s really interested
in deep learning and he’s confident that with my drive and
passion that he spotted, that I’m going to be able to give him
a lot of input on that. And I’m going to do my best to do so,
I’m confident I will. Yeah, I want to really dive deep into it.
Kirill: Nice. I totally agree with that. I’m sure you’ll be able to help
him along the way as you learn these things yourself. But
also for me, with my mentors, I wouldn’t be where I am
without my mentors. So if at any point they call me and say,
“Hey, Kirill, I need your help with this,” I’ll just pause all my
plans, I’ll go there and spend a week or two weeks helping
them implement the system or whatever I can do, set up
some contracts, or introduce somebody to somebody, or
even if it’s something like deep learning AI, I will make sure I
do that. I’m sure you’ll do the same.
If five years from now, when maybe he’s not your mentor any
more, he says, “Hey Rico, you’re the CEO of Deep Learning
Incorporated Global Worldwide. Could you come over and
help me?” (Laughs) Richard, this is where you grow some
lettuce. And I’m sure you’ll say yes.
Rico: Absolutely, yeah.
Kirill: And I like how you mention—this is so funny. I’ll show you
my book. This is the top of my book, I’m making notes right
here. This is before our talk. I wrote down ‘passion’ and
‘drive.’ You mentioned those exact two things that Richard
spotted in you. I was like, “When I talk to Rico,” this is what
I wrote down I need to talk about. First two things that came
to mind – passion and drive. This is cool.
It shows something and it means that you emanate these
qualities and people can pick them up. So, let’s talk more
about that. What is your passion, and seeing what you’ve
seen in computer science, where in the space of AI do you
want to go and why are you so passionate about it? And why
are you so driven to get there?
Rico: Where I want to go, I want to develop amazing things.
Actually I’m really inspired by Ben Taylor right now because
he posts a lot of his stuff on LinkedIn. He had a five-article
series about how deep learning is used to spot beauty in
people, but then filtering out the race aspect.
Kirill: Oh, okay, how deep learning is used to prevent racism, to
combat racism in beauty, fashion stuff and things like that.
Right?
Rico: Exactly. And that’s the stuff I really want to get into,
because I think—also I saw this TED Talk the other day
which was about a guy who created a drone system in Africa
where they have autonomous drones delivering blood to
different hospitals, which is saving lives every day.
These are the types of applications that I’m really interested
in, that I want to create, because I think that AI can not only
automate, so augment humans in what we do in our
everyday life, it can also make the world a better place by
applications like this. That’s what keeps me going because
I’m always looking for—everyone is looking for his purpose,
right? Everyone is in his world trying to find his purpose and
what he can contribute to the world. And I think I really
found that for me in artificial intelligence. That’s what keeps
me going because I think it has relevance.
Kirill: Man, everybody is looking for their purpose, but not at 20
years old. You’re way ahead of the curve. That’s crazy. How
did you get there at such a young age? It’s very inspiring,
but it’s a mystery. A lot of people at your age are still
searching, or taking a gap year, or partying a lot. Was there
something in your life that changed your perception?
Rico: My year abroad.
Kirill: Dishwashing in Canada?
Rico: (Laughs) Absolutely. I think that would be my number one
recommendation for everyone who is in a young age, who is
getting out of school, because we have to start going to
university at some point. And it was young students who
were wondering about doing the same co-op program instead
of a normal study. And they were wondering because the co-
op program is more intense because you really never have
holidays, you are always working or are in school.
So what they were wondering, they were 18 years old and
they said they would have a disadvantage if they don’t go
into university right away. I would say, “Guys, this is not
true.” Because after school, you kind of want to enjoy, live
life, so that’s what I did with going to Canada, and I think
that’s how I built up street smartness. And I feel like that’s
also important. I’m getting lost on this one. (Laughs)
Kirill: (Laughs) So what happened in Canada? Why did you come
back from Canada and you’re like—oh, that’s right, you said
you wanted to get into the movie space.
Rico: Yeah.
Kirill: And then there, through the whole coincidence, computer
science was one of the predominant things. That’s how you
ended up in AI. That’s really cool. You know, Steve Jobs
talked about connecting the dots. Before you went to
Canada, you never would have thought that these
coincidences, that’s where they’re going to lead to. But now
looking back, you can see how connecting those dots makes
sense, like this had to happen for this to happen. That’s
really cool.
Okay, and then the other thing is—tell us about your
transition, the way you’re entering this field. It’s obviously
not one of the simplest fields in the world. You know how
sometimes they say – no offence to any cooks in the
audience – that cooking is not rocket science? I think it is
rocket science, I’m not a good chef, but there’s a saying it’s
not rocket science. Well, you can’t really say that about AI.
AI is pretty much rocket science, it’s like almost there. So
you’re getting into one of the most complex areas in the
world and the most cutting edge technologies and data-
driven applications. How does it feel? Like, what are the
challenges that you face on a daily basis?
Rico: The point you just mentioned is the thing that’s really
getting me excited about AI, the challenge behind it. That it
is something that is not yet developed and there’s still room
for us to improve. So I really like that challenge in the first
sense. And I think the biggest challenge for me right now is
to incorporate my passion for AI with my daily study life.
Because I still want to finish university, I still want to do
great on my Bachelor’s, but my computer science degree is
not that connected to AI. So there’s separation between my
Bachelor’s and then also my AI passion. And kind of getting
this under the same hat right now is my challenge.
Kirill: That’s a really cool way of putting it, getting them under the
same hat. I wanted to ask you, is time a challenge? Do you
have enough time to do your Bachelor’s, your work, and
your passion for AI? I’m kind of leading towards your whole
sleeping routine. I would like to talk about that and how
you’ve changed your sleeping routine since DataScience GO.
Rico: Inspiration from Hadelin. You want to control time rather
than undergoing it. That really stuck in my head because it’s
true. Ever since I started the sleeping rhythm, I have much
more control over my time. Before, I’ve been stressed, so I’ve
had these days where I really wanted to get stuff done and
then I was stressed at the end of the day because I didn’t
have enough time to do so and then I was not happy with
myself. Well, now, obviously I’m not always on top of my
game because it’s also been a huge change to get into this
habit, but I have all the time in the world now. I think it’s
good for me because I’m not stressed anymore and I know I
can get things done during the day and I do get things done.
And that keeps me motivated. That’s just been my
motivation to stay with this.
Kirill: So, tell us a bit more about that. How many hours a day do
you sleep now?
Rico: I do sleep about four and a half hours every night. And then
I do two 20-minute naps over the course of the day. So
basically I split my day into three parts. I wake up usually at
5:00 or 4:30, and then six and a half hours later I have my
first nap, and then six and a half hours later I have my
second nap, and then six and a half hours later I go to sleep.
So splitting my day into three parts is really cool because
sometimes when you wake up you’re really energized and
you’re ready to get going again – I get that three times a day.
Kirill: Nice. That’s really cool.
Rico: That’s really getting me going.
Kirill: Was it hard to transition? What was the biggest feeling or
complication that you felt while you were transitioning?
What was the hardest thing?
Rico: It was hard and it still is hard. I can’t lie, getting up in the
morning sometimes is pretty hard. That’s also what Hadelin
told me: Sometimes you have to break the habit, just sleep
in for a day. For example, after my first job, where I was
mentally ready to get some rest, I just slept in for like 12
hours. And that was really good, because the next morning I
woke up being more energized than ever.
Getting up in the morning is sometimes hard, and I think, to
anyone who has been thinking about doing it, you have to
fill your day with work. You can’t do it if you don’t have a
full day of work in front of you. Because once you feel like
you’re wasting your time with non-related stuff, you might
as well sleep longer. I try to use my time as good as possible,
get the most out of the day, and I think you also have to be
excited about your work, because you cannot get up at 4:30
when everyone is sleeping if you don’t like what you’re doing.
Kirill: Exactly. I was just sitting here thinking I have to mention
this for our listeners because so many people listening to
this podcast are expressing concerns about the health of the
guests because even to me, it’s starting to feel like we’ve got
a cult going on here. Like, Ben Taylor sleeps 4 hours a night.
You sleep 4.5 hours a night. Hadelin sleeps 3 hours a night.
I tried sleeping 4 or 5 hours a night.
It’s crazy, the amount of people that I interact with through
the podcast, through the conference, the students, it’s just
surprising how many people are doing these routines, which
some are called polyphasic sleep cycle, Uberman sleep
pattern, there’s many different versions of it. But what I
wanted to say is that it’s not a cult and it’s exactly what you
said. When you’re passionate about what you do, you can’t
wait. You’re like, “Why am I sleeping? I have to get up and
do more. I’ve got to get back into it.”
Of course you have to do it very consciously and monitor
your health and take care of things, but for people who do
master it, it helps. And not only with work. Like, in your
case it helps with work and study, in Ben Taylor’s case, he
gets to spend every evening from 5:00 P.M. to whenever he
goes to bed, around 10:00 or 9:00, he spends it with his
family. It helps find the time and still get those things done
or like you were saying, put several things under one hat.
Anyway, we veered off a little bit from your AI and your
passions. Okay, so that’s what inspires you in AI. And the
challenge inspires you as well. So, tell us how do you go
about learning AI. You mentioned you took a few courses, so
you already have the foundation of what artificial
intelligence is. Do you go and decide to code your own
neural network for some sort of application to practice or
you find a real-world challenge and you try to solve it or you
find a dataset and you want to get some insights into it or
you just apply it at work? Like, what is your way of getting
more than you’ve already gotten from the courses as your
foundation? How do you propel your skills in the space of
artificial intelligence?
Rico: That question really addresses my problem right now. As I’m
studying, I don’t really have the time to dive deep into the
application of models. So for me it’s like right now, I’m really
getting into the theoretical, like the part that really interests
me. For example, the future of AI is something that really
sparks my interest. I try to listen to a TED Talk every day
that’s relevant to AI. I do the courses. I connect with
LinkedIn so I have a great feed now that always keeps me
updated on the newest technology and the newest methods.
And then once February starts, after my exams, I’m going to
go ahead knees-deep into everything and I’m going to do
practical applications. My plan is to really get a github
account going in the two months I have and get Kaggle—you
heard about Kaggle?
Kirill: Yeah, I was just thinking. That’s the best place to apply your
skills.
Rico: Yeah, definitely that. And real-world problems. So what we
also try to do in the research lab is we have a bunch of
customers in our company, but also a bunch of products
that we can integrate AI into. So, we want to keep the
research phase really slim, so after two or three weeks when
everyone has gotten into the topic, we want to address
company’s problems, company’s issues, and try to improve
our product with AI. And that’s where I’m really going to get
the practical knowledge. And I’m really excited for that. I
actually can’t wait. That’s what I meant. Like, I have to stay
committed to school, I have to finish my exams first, and
then AI comes second in that sense.
Kirill: Nice.
Rico: But yeah, I have everything planned out.
Kirill: I like that. I like how you have things planned out. Even to
the date, right? And that’s the difference between goals and
dreams. Dream is when you’re like, “I want to do that,” but
you don’t know when. Goal is a dream that has a timeline,
like, “I will do this in February and then by April I’ll have
this experience and so on and then I’ll start my next
semester, etc.” What I wanted to ask you is trends,
technological trends in AI, you’re excited about them. What
are some of the top ones that you’re most excited about,
something that you think is going to happen in 2018?
Rico: Capsule networks.
Kirill: Capsule networks?
Rico: Yes. Ben Taylor is a really inspiring person, by the way. Ben
shared an article on capsule networks, so I watched a little
video on it which really sparked my interest, so I’m going to
have a seminar next semester which I’m going to prepare
starting January, which is basically a 60-minute
presentation about an AI-related topic. I asked my prof if I
can do it about capsule networks, because that’s something
that I think is going to be relevant in 2018 so I really wanted
to get deeper in the knowledge about that one. And I’m not
actually sure because I haven’t done too much research on
the blockchain, but I think AI BlockChain is going to be a
trend for 2018 like he has also mentioned yesterday. I’m
really interested about that as well.
Kirill: Nice. That’s really cool. How did Ben Taylor phrase—I really
like the way that you go about things. “I don’t know
anything about capsule networks, but I’m already presenting
on it in April.” And that’s really cool. And Ben Taylor put a
phrase to it. It’s like some sort of commitment, radical
commitment. Do you remember what he said? It was like—
Rico: Reckless commitment?
Kirill: Reckless commitment! There we go. That’s exactly it. You’re
living up to what he was preaching about. Reckless
commitment, that’s the way to get things done.
Rico: Yeah, exactly. Maybe I can try and explain why I’m such a
fan of Ben in that sense. I think on the first day after his
talk, I approached him and asked him if he wanted to have
breakfast with me the next day. So he said, “Yeah, let’s do
it.” So we met up the next morning, had breakfast together,
and I was able to ask him all of my questions concerning
how I can approach management about my idea for the AI
research lab, what his opinion is, how he would pursue that
presentation.
And he really gave me amazing input about everything. He
gave me two or three techniques about how I can approach
management with it and we had a really nice chat about his
applications, what he’s been doing, and where he’s been
coming from and that’s why I’m—really, he gave me great
input on DataScience GO. Definitely one of my top interests
there.
Kirill: That’s so cool. And Ben is like an ocean of stories and advice
and crazy things. Like, you can talk to him for hours. We
were talking at dinner and he just mentioned story after
story after story from his life, it’s crazy. And I actually
remember that breakfast, I saw you guys having a chat
there, so that’s really cool. It’s really cool that he gave you
some input to help you out with the AI lab. Was it hard?
What was the biggest challenge in approaching your
company? When you said, “I talked to my company to set up
an AI lab,” I was like, “What? Who does that?” Who goes up
to their executives and says, “Hey guys, we need to set up an
AI lab?” By the way, you haven’t told us yet, what does your
company do? That will probably give us a bit more
perspective on how hard it was to set up an AI lab.
Rico: Definitely. We’re an e-commerce ready company. For
example, for products we have an online shop software and
a product information management system, and we also sell
online shops to huge customers and we do customer service
software. So, yeah, it’s all e-commerce-related and we’re a
really customer-focused B2B company, so we had AI
integrated in our company before with the chatbots, and we
have an intelligent mail system.
But then I thought, I looked at the e-commerce use cases for
AI, also something I asked Ben, I was like, “Ben, is there
even applications in e-commerce?” and he was like, “There’s
plenty.” So I googled it up, I found so many and I was like,
“Okay, let’s do this.” So I presented them to management
that has major, huge cases – maybe I can talk about that
later – and we can definitely use them to improve our
products, to improve our customer service, and also grow as
a company. So, yeah, apparently my management was
already kind of thinking about integrating AI, so it came
really perfect for them in timing. And they’re really
supportive with it, which is nice.
Kirill: That’s really nice. And then they were like, “Okay, Rico. You
take care of it, set up the lab.” What does it entail, setting up
an AI lab at a company? What does that even mean?
Rico: We’re in the process of building it right now. We chose a
team building approach, so there’s going to be a meeting
tomorrow, actually, where there’s going to be interest, like
students coming to our company, and we present them with
an idea, and hopefully they’re interested to help us build it
up. So, we want to acquire a team, like 4-5 people, and then
define a structure of what we actually want to do, which is
integrating AI into our products first, do a proof of concept
that it actually works, that it makes our products better,
then talk to customers about it, approach customers and
say, “Hey, we’re going to integrate AI into your products,
maybe analyse the data with just data science lessons, and
you will have x improvement and it will save you x cost.”
That’s also something we’re going to work on.
Yeah, just like general integrating. I’m sorry, I’m drifting off,
but you also said yesterday that 40% of companies already
adapted to AI. I think it’s really important for companies
these days to get knowledge in that field, to not stay behind,
so I feel like an AI research lab is perfect because we are
going to do research and we are going to provide information
that can be really useful for the future.
Kirill: That’s awesome, I love that. And for those listening, if you’re
in a B2B space, that’s a great place to apply AI. Because for
individual people it’s a bit harder to make the connection or
explain how AI is going to benefit them. If you create an app
or something that massively is going to help people, like lots
of people, that’s cool. But overall, if a company comes to me
and says, “Hey, we will build you this AI thing,” as a person,
I won’t be able to pay them tens of thousands of dollars for
an AI application for me, so you have to focus on apps and
stuff.
But in the B2B space, because there’s much higher
turnovers and much higher funds that these companies
have, like in your case, Rico, you can just come up to a
company and say, “Hey, we know that you can increase your
efficiency and we know that that’s going to cut your costs by
10%.” And if their costs are like $10 million, that saves them
$1 million, and you charge them $100,000. It’s a no-brainer
for most businesses.
And that’s why the saying “AI is the new electricity," which is
by Andrew Ng, it’s so much deeper than people think. It’s
not just about that AI is going to be everywhere. It’s also
about how quickly a company is going to adopt that, and
how easily, if you show them the bottom line. “How is that
going to change the bottom line? If it’s going to save you a $1
million and you just have to pay $100,000, where is the
question? Let’s just do it.” You know, I can totally see how
that can be beneficial in your case.
All right, e-commerce use cases. We’d love to hear some of
those if you can share or remember any of the ones that
popped up when you were searching on Google.
Rico: Yeah, I was going to do the three basic ones that I found the
most information about. Chatbots on an online shop, like a
chatbot that you can ask. For example, if you’re on some
shopping website for clothes, you can ask the chatbot, like,
“Show me your black shirts,” or, “Show me something that I
can wear on a Saturday night.” It’s an intelligent bot that
helps you navigate through a website and find products.
But we’re already I think into that space, so I think what’s
more interesting is personalization of the customer
experience, personalizing the shopping experience, and also
the sales cycle. Because once you create a user profile based
on their data, like how much time they spend on the
products or where they click from, you can use clustering to
cluster your customer base and approach different segments
of customers in a different way. And that can be beneficial
for sales.
And then also apply NLP and image recognition for
intelligent searches. So, using keyword mapping, if I put a
keyword in and it doesn’t only show me the relevant results,
but also related results based on the machine learning
algorithm that compared the words. Or if I think Kirill has a
great shirt, I’m going to take a picture of it, upload it into the
shop, and then the shop will not only show me Kirill’s shirt,
but also other shits that may be similar. And also what I
think was not necessarily e-commerce related, but
something for our company is in the space of ops, where you
can use machine learning to analyse backlog data to have a
proactive error detection. That’s also like the main thing I
found out.
Actually, when I had a meeting, when I presented it to my
management team, one of the guys, the head of the product
information management system, he was really into it and
he was like, “Oh, I have so many use cases we can talk
about.” Yeah, so I’m excited to hear about that and there’s
definitely a lot of stuff, a lot of projects for us to work out.
Kirill: Yeah, that’s so cool. It sounds like there’s so many that as
soon as you get started, you’ll have the problem of “Where
do I get the people? How do I do all these things at the same
time? It’s just impossible!” Are you a bit worried about that
already, or are you just excited about it?
Rico: I’m really excited. I actually can’t wait to start. It’s kind of
bad because I have exams in February. I really want to slay
on that, but I’m really excited about what’s after. I can’t wait
for mid-February to start it.
Kirill: Gotcha. So, one question I have is: data science versus AI.
What would you say to those listening? Because this is a
podcast for data science careers, what would you say to
those listening who are in the space of data science, and
quite successful and learning in that area, but they’re a bit
apprehensive about getting into the space? It seems like a
whole different area, something to do with development,
something to do with robots, and it just feels very alien to
them. What would you say about can they do it and should
they do it?
Rico: I feel like, because you already mentioned, that machine
learning is going to be a great component in the future for
data scientists. Therefore, deep learning, which is a
subspace of machine learning, and deep learning is what I
mainly talk about in my applications, they will all be solved
with machine learning or deep learning algorithms, I think
it’s going to be really important for the future. And I think
they’re definitely not going to regret getting into it because
it’s just going to provide a way to get information out of the
data with algorithms. And a programming language such as
Python is not that hard, so I think every data scientist
should really have a look. Well, if they’re not passionate
about it, it’s not a huge deal. But if they are, it would be
great for them to try it.
Kirill: Yeah. How long did it take you to learn Python?
Rico: Well, I know Java from school, so I took me like four days.
Kirill: Four days? Okay.
Rico: Because it’s the same as object-oriented programming,
right?
Kirill: Yeah.
Rico: But in general, I think Python is really simple with how they
structure the data and also the variable types, so I think it’s
really a great programming language for everyone to get
started.
Kirill: And the syntax is very simple, right, with the whole spacing.
It’s one of the easiest ones to learn ever.
Rico: Yeah.
Kirill: And do you use TensorFlow, or do you use PyTorch? What
do you focus on mostly?
Rico: Well, I got insight of both of them over your courses. I think I
would prefer TensorFlow for no specific reason, but that’s
probably what I’m going with. Just for me, the general
feeling, that was more comfortable.
Kirill: Yeah. And I think it’s more templated. It’s very easy. And
that was actually my question. Do you feel that with
TensorFlow and PyTorch—but in this case with TensorFlow,
you can apply sophisticated AI models without actually
coding a lot of lines of code? Like, ten lines of code and you
could have a convolutional neural network set up, right? Do
you find that makes your life easier?
Rico: Yeah, especially when you put Keras on top of TensorFlow,
it’s even easier.
Kirill: Yeah, exactly.
Rico: But, yeah, I think that’s a great way to get started. So for
me, when I was getting started, I was really happy with the
results that you could accomplish really easily. And then I
feel like that’s a really good entry into the field, because once
you start seeing the results, you’re happy. And then you’re
interested and you want to find out more about the
algorithms. And then you dive deeper into the mathematics,
maybe read a paper about it, and that’s when you really
start going. But I think the possibility of Keras and
TensorFlow really sparks the interest in the first place.
Because as a computer scientist, I know it can be really
depressing when you do a lot of work, but don’t get results.
In that sense you get the results first, and then you’re
inspired to do the work, at least for me.
Kirill: Yeah. That’s really cool. And the applications, even the
practice applications—in AI, when you’re learning AI, I think
across the board, regardless of which course you do, I think
they’re always fun. It’s like, you’re trying to recognize digits,
or you’re trying to classify dogs and cats, or price properties
and stuff like that. They’re always really fun applications. In
machine learning, unfortunately, there are some more
historical datasets like Virginica, Setosa, Fisher’s Iris dataset
and so on, that are kind of very textbook.
But because AI is so new, most of the applications that you
can see online and most of the tutorials, they are really fun.
And when you see the results, it’s like, “Wow! That is a dog.
That is a cat. That is so cool.” I love that part as well. Okay,
Rico, I don’t know how quickly this hour flew by, I’m not
even keeping track. I think it’s been an hour, but it feels like
5 minutes, it’s been amazing. I had a question about a book.
Do you have a book that you can recommend to our
listeners to help inspire them?
Rico: Yes, absolutely. The book I read that also got me really
excited was “The Magic of Thinking Big” by David Schwartz.
Kirill: Yeah, yeah. Is that the one with the fish on the cover, right?
Rico: No, that’s “The Big Leap.”
Kirill: That’s “The Big Leap,” okay. I haven’t read “The Magic” one.
Okay.
Rico: Yeah, “The Magic of Thinking Big” is a really great book
because it encourages you to dream big. Because when
you’re not in a position when you have achieved a lot, which
is in my position right now, you have to dream big in order
to achieve big. Because when you see yourself in that light of
where you want to be, you will behave more successfully and
you will be more successful. And also what really inspired
me by that book was when he quoted “Successful people are
always going to appreciate big ideas.” So that’s what really
got me with my company’s presentation, because when I was
thinking about if they’re going to like it or not, I was like,
“Well, it’s a pretty big idea, so they will appreciate it.” And
that’s what happened. And that’s what really got me to
making a presentation. And if you allow, I actually want to
recommend another book.
Kirill: Yeah.
Rico: “What the CEO Wants You to Know.”
Kirill: Sorry, repeat that?
Rico: “What the CEO Wants You to Know.”
Kirill: Oh, “What the CEO Wants You to Know,” okay.
Rico: It’s been recommended on the podcast before. I just wanted
to reiterate that because I’m reading it right now. It’s great
for anyone who wants to get basic knowledge about what
their business is doing.
Kirill: So it’s not necessarily for CEOs, it’s for anybody in the
business?
Rico: Yeah.
Kirill: Okay, cool. That’s a good recommendation. “What the CEO
Wants You to Know.” And actually why that book resonated,
“The Magic of Thinking Big,” I actually read it a few years
ago. I looked up the cover and it’s this white book with big
red writing by David Schwartz. It’s a very old book, it’s
written in the ‘60s or earlier, but very true. That’s really
impacted me as well, so I can vouch for that. That’s a great
recommendation.
Okay, Rico, thank you so much for coming on the show.
How can our listeners contact you? What’s the best way to
get in touch and see what crazy AI applications you are
going to create in the coming years?
Rico: LinkedIn.
Kirill: LinkedIn?
Rico: Definitely. I love to connect with people on LinkedIn. I also
love to go out and talk to people on LinkedIn. I’ve gotten
some great inputs. I also looked for speakers for my Meetup
on LinkedIn. I think LinkedIn is right now the greatest way
to connect with people also in the business. Yeah, definitely.
As soon as I will dive deeper into everything, I will also try to
do more AI-related posts.
Kirill: Nice. And your Meetup, is it going to happen again, or was it
just like a one-off thing for now?
Rico: It’s going to happen again.
Kirill: Awesome! And which city is that in?
Rico: Hamburg.
Kirill: Hamburg. So, if anybody is in Hamburg watching this or
listening to this, make sure to check out the Meetup. What
is it called, meetup.ai?
Rico: It’s on meetup.com, that’s the website, and the Meetup is
called meetup.ai.
Kirill: Meetup.ai in Hamburg, check it out.
Rico: We can put the link in the description.
Kirill: Yeah, yeah, we’ll definitely put the link in description, but if
people for some reason forget to check the description, make
sure to check out meetup.ai if you’re in Hamburg and go
meet Rico in person, get inspired, get some of his energy.
Rico, thank you so much for coming on the show. This has
been crazy amazing. I’m sure so many people are going to
get inspired and energized by everything you shared. Thank
you so much.
Rico: Thank you so much for having me. It was great fun,
actually. (Laughs)
Kirill: All right, so there you have it. That was Rico Meinl on
artificial intelligence and his journey into data science. I
hope you were inspired, I hope you got that energy. I
definitely felt the energy from Rico and just how powerful his
ambitions are and how powerful his drive and passion are.
It’s just incredible to meet people like that who set
themselves some crazy commitment and they just go forward
with it.
Personally for me, that was the biggest takeaway, that the
reckless commitment concept works. If you set yourself a
goal, if you set yourself a target and you just commit to it
and you know that there’s no way out, you have to go for it.
As they say, “If you want to take the island, burn the ships.”
There is no way out, there is no turning back, you have to do
it. And the way to get that ‘no turning back’ is you promise
someone or you talk to someone and you say, “Hey, I want to
do this,” or you set up a Meetup group and you know there’s
250 people waiting to come to the event and you cannot let
them down, or you talk to your executives and you say, “I’m
setting up this AI department, it’s going to happen, and
there’s no way, no turning back, no way back.”
So there we go, that’s the biggest thing I’ve learned. I would
love to hear what’s the biggest thing that you learned. And if
you enjoyed this podcast, make sure to rate it on iTunes and
really help us spread the word across the world so that more
and more people can get amazing insights like this. And of
course, you can get all of the show notes including the
LinkedIn URL for Rico’s profile and also the Meetup group.
You can get that at www.superdatascience.com/123. Once
again, this episode is available in video, so if you just listen
to the audio, you can maybe later on someday go back and
re-watch it in video to get inspired again and get a different
experience. That is also available at
www.superdatascience.com/123. And I can’t wait to see you
next time. Good luck with your reckless commitment. Until
then, happy analysing.