So I think that if you give you
definitely will get back in different other forms, but it brings
lots of opportunities for everyone That's definitely my thought about it Hi Amel, how are you doing today? Good! How are you, Rebecca? I'm really good, thanks! I'll just explain that I'm doing
series of podcast speaking about women and AI and I get to speak to some of the Great
engineers and researchers that we've got at InstaDeep
and I'm really thankful to you for volunteering to chat with me today,
because I know that you're super busy You're not only a machine
learning researcher, an engineer, but you're also a team lead
and a project manager So I'm really glad
that you've been able to fit me in today Thank you for for the invitation Really glad to be part of the podcast what I find quite interesting with you
is that you've been in InstaDeep for over five years and there's not many people
who can say that So a lot of us have only been here
for a couple of years and InstaDeep is ten years
old this year So that means you've been here
for more than half of its life and you've experienced
all of the big milestones How is that being to experience
so early in your career? Yeah, I mean, like the opportunity
is incredible here at InstaDeep I've worked in different
kind of projects Met lots of incredible people to learn from I mean, the ups and downs
from all the early stages from how to go from the series A
to series B, and then the acquisition was really actually a very exciting time,
and especially generally when we speak about the acquisition,
how we go from an early stage startup basically
trying or going through the everyday work from working as a machine
learning engineer, trying to also share in the community in different
kind of events with InstaDeep I think the whole journey
was really incredible That's really cool. So how many people were there
when you first joined InstaDeep? Basically When I joined,
we were like 30 people. going from a small team, like knowing pra
ctically everyone
to growing to more than 200 person or people in the whole company
knowing some, not knowing others and getting to know the different culture, the different kind of ways of how people do things
in machine learning. or AI is really interesting. That's a good point that you put up there the differences in geographies and
cultures that we've got within InstaDeep I'm in Paris,
you're in Tunis We don't work on the same team, but we do have teams
where people are in many different pla
ces And thanks to modern day technology,
we can work really well like that Do you find it adds
to what InstaDeep can offer? Yeah, definitely I think basically
working in different cultures, learning from different other people,
working with the amazing people for example, here from Tunisia,
we're working with people that have studied in different big schools
like for example Oxford and UCL that is a really incredible
opportunity to be honest. So meeting these kind of people
who really push it in
the machine learning industry
and you get the chance to learn from them. It's really an exciting opportunity
to be honest. that's really cool. And now people learn from you and all of your experience.
Maybe we could go back in time a little bit I know that you studied computer science
originally and then you said that there was something that sparked
your interest in machine learning what was and why did you think That that's really something I want to work in Basically, when when I was doing
m
y computer science engineering degree there was a kind of event at my university at that time and people came and talked
about machine learning, talked about data science I was like "wow"
this is really interesting Basically you try to predict or you tried to mimic
the human nature and try to create something that resembles the human nature
or the human being themselves So I thought "how do they think about these things" I thought this was really something incredible and something really
interes
ting to be honest I'm really very fascinated
about all this field and how it incorporates lots of maths in it and how we create from like this huge math formula, We create something that tend to think
Like the human being, which is incredible. And the journey started there. So I really wanted to work in something
related to machine learning. All my internships were
towards machine learning, even though in Tunisia, But back at that time, there were no People or companies
who worked in this kind o
f field. it was really incredible. So you mentioned
the Google developer experts and I was looking this up the other day because I saw that you were
a google developer expert I've seen it's just anyone
that they accept as a Google developer exper but you have to show
that you've got an expertise your subject So how did you go about that and what kind of things you do
as a Google developer expert? I had this opportunity of being a Google developer expert
but like as part of InstaDeep and I think
like it started from,
from a very basic need for me, which is basically helping people to learn
more about machine learning I was thinking about just being a Google
developer expert or something like that then things happened
and there is a kind of process about it when you when you share a lot,
when you give like people, like more opportunities
to learn about machine learning and things like that,
this is how I became one I was like very active at that time. I had to do more ..., for example, s
ometimes I can even do, for whole weekend,
two workshops at the same time, one on Saturday,
one on Sunday just to give more people the opportunity
to learn about machine learning this was part of
the Google developer expert program And the goal
as everyone knows, is to share knowledge related to a Google technology
can be in machine learning. It can be, for example in cloud
or in any other kind of things. So you mentioned that you're trained
to become a Google developer expert because you were k
ind of already doing
the things that were required to be one anyway so sharing this knowledge
and that was really important to you. Yeah. you also mentioned that
like giving back to your community is important to you
and that this is one way that you do that How important do you think that is for
people in machine learning to do I think as everyone says
"alone you go faster together we go far" and if you bring lots of people
that share a lot about machine learning. You bring more opportunity
for
for everyone in this way Because if you do everything alone
and you have like that kind of mindset, of saying
"No, I don't want to share this, this is my information." So basically other people
won't have that kind of giving back to you when you give them. basically if you share, more people
share, the circle gets wider and wider So it brings more opportunity for everyone More companies will start
working on machine learning. When we talk about Tunisia this can help to have more companies based
in Tunisia working on machine learning
having more companies who are well known in Tunisia If you give, you
definitely will get back in different other forums, but it brings
lots of more opportunity for for everyone. That's definitely my thought about it.
Yeah. this is something that I sort of noticed
when I moved from more traditional fields towards machine learning was
how many things where open source How many publications are like
freely available Whereas in the fields that I've worked in
i
nthe past there was a paywall behind,
to read an article. So if you're in a university
or you're in a big company that can afford to pay that,
then you've got loads of opportunities. But if you're just somebody at home
who has a computer, who wants to learn about these things,
then you're blocked And I think there's a lot fewer barriers in machine learning for that. And it is because we've got people like you
in the community who are really passionate about, making sure that things are open sour
ce,
that knowledge is shared and this is accessible
to as many people as possible. So that's really great. Yeah, definitely. I totally agree on having, in the machine learning community, it's completely different from,
other communities where people really want to share about all the advances that is happening I think this is how machine learning
was brought together. if for example, the machine
learning community was not like that, this kind of open source. I don't think that in this way
we wou
ld be what we are right now or having all this kind of models
or this kind of technologies or this kind of advances
from to chatgpt, from everything. What would you advise people
who want to be doing the same job as you? What would you advise them to do? So basically,
I would say of course, work, work, work. a lot of work. If you really want to work in machine learning and you really love it
I think passion is important if you don't think like machine learning
is exciting, you have to find The t
hing that you really love doing
can be in any kind of sector, actually. But in machine learning,
what I really recommend is that if you really want to be one of
the good machine learning engineers, apart like from doing like project,
which is a very important thing You have to do the research
or let's say understanding the mathematical
behind how machine learning works. Because that way
you'll be able to understand, like have new ideas and come up with solutions
to the problem that you have. Oth
erwise you will be stuck somehow
not understanding what other people are talking about or what other people
are saying or planning to do And you wouldn't have that
kind of background that will help you in solving the problem that you have because machine learning is very somehow
stochastic at a certain point And if you don't know exactly
what should be done if you don't have lots of data,
what is the thing that you should do the kind of model, the choice of the model,
the choice of the parameter
s and how you engineer your way around it
is very important And to be able to do this,
you have to read and be up to date, with the research papers
with what's happening know it's very hard to do so
and it takes lots of time and momentum to be able to read machine learning papers as they say "practice makes perfect" So you start slowly Everyone start from somewhere small
and then, try to make it biggera and bigger every time
so this is what I would really recommend. And networking is also impor
tant As I always say
because this is where meet people who share with you opportunities
or share with you new ideas Okay. So I guess make sure that you've got the basics
right make sure you understand the math. and keep up to date to reading research papers and I guess that's one of those things that gets easier as time goes on, right? The more you read,
the quicker you'll be able to read the next paper because you already know the background. So yeah, exactly.
It's an exponential curve in that
way So like you said, it starts off difficult,
but it doesn't get super easy but it does get easier
as time goes on. so that's being very good
on the maths and the technical side of things you said working on projects So that means practical experience
coding I guess Yeah, definitely. if you don't have like
practical experience, it would be very hard for
you even to work with, other engineers
who have like more experience. So having like the basic knowledge about
how applications work, how machi
ne learning works and the intuition behind it
is also very important. Okay. Yeah. So I guess there's
a lot of resources online if you are already kind of in the field
where they can get advice on this. And we mentioned before,
open source projects getting involved in those, is a good way
of getting experience as well. So yeah, we could, we could speak
for a long time on all of these subjects. But we also have have work and coding to do. So is there anything that you feel
that we've not spoken ab
out that you would like to say before we go though, I think we covered
lots of topics and good luck for everyone who wants to be
a machine learning engineer I really recommend you put a lot of effort and work in it I know it might be very frustrating
at the beginning, but at the end it is definitely going
to be a rewarding, hopefully amazing. Thank you so much Amel for your time and
see you soon. Thank you very much Rebecca Thank you.
Comments