YouTube, 4x Google and OpenMined. Meet Irina.
Irina started coding at 10. Since then, she landed multiple internships at YouTube & Google (even in the US!). She did her Master thesis at Google DeepMind and currently is a tech lead at OpenMined.
Studies: MSc Data Science @ EPFL ‘20-’23, BSc CS @ Alexandru Ioan Cuza University ‘17-’20
Experiences: Tech Lead @ OpenMined ‘23 - Present, Research @ Google DeepMind ‘23, SWE Intern @ Google ‘21, SWE Intern @ Google ‘20, SWE Intern @ Google ‘19, STEP Intern @ YouTube ‘18
Associations: Startup Manager @ EPFL-UNIL Entrepreneur Club ‘21-’22, Managing Partner @ Founderful Campus ‘21 - ‘22, Managing Director @ Girls Who Code ‘17-’20
Origin: Romania
Links: LinkedIn
Looking for jobs/internships at exceptional companies?
I just launched the Ambitious x Driven jobs board!
These are companies that I personally screen for exceptional founding teams, who have raised funding and where you can learn immensely :))
🌐 Fabric Building user-context as an API - hiring AI & Full-Stack Engineers
🏢 AssetOS Transforming real estate due diligence using AI - hiring Full-Stack
🤖 Rapidata Near-instant AI data annotation used by e.g. IBM - Founder’s Associate
🏥 Alpina+Sana Solving hospital malnutrition using CV - CV Research, Software
Growing up in Romania & working on side-projects
Where did you grow up and what were you doing?
In Romania, I discovered programming at age 10 through free local courses. Programming itself was interesting, but I got really fascinated by electronics - for a couple years, I built line followers, sumo robots and it was fun! I even learnt many physics concepts before they were taught in school.
Like many Romanian students, I did competitive programming for 7-8 years and that was very formative for my thinking. My interests eventually shifted from algorithmic to more realistic development, so I ended up participating in software contests where you present whatever you’ve built (games, apps, websites). These were great because I saw fellow high schoolers building Unity games with thousands of downloads and it fuelled my passion to build things in my free time.
I started with Unity and found myself some wonderful mentors in indie game development. It was an interesting community because everyone was building purely out of passion. This led to my first job at 16: building middleware for Unity where I worked on a memory pooling toolkit. It was tough to balance high school with work, so I stopped after a while, but not before we built my hometown’s first VR app for tourists - quite an achievement at that time.
Was this normal for your age? If not, what made you pursue these extracurriculars?
No, I was quite unusual and had my own path. I was fortunate to experience many different communities, and in each one, to find people who were truly passionate about their craft and had a contagious energy.
This continuously inspired me. To be honest, attending national olympiads in Romania also meant getting a lot of free time off from high school. As olympiads were approaching, you could skip regular classes and only attend math, literature and physics. This flexibility gave me the time to learn and build out a lot of my own projects.
You then started your bachelors in Romania. After your first year, you already landed a Google internship, which is super impressive! How did you even think about applying? Did you have mentors that suggested this?
During my first year at university, I was already working full-time and I was happy. On our way to the ICPC ACM regional competition, our trainer offered to recommended me for a job at Google. I knew little about Google and was not particularly interested in joining.
I still took the interview, but didn't take it very seriously, rather for self-evaluation. Given that, I was very relaxed and I ended up getting the offer. My parents were upset at the idea of me not going, so I ended up going. It turned out to be really fun - going abroad, meeting people. My first internship was in Switzerland and I got a taste for it. Ended up doing 4 more internships at Google after that.
It’s interesting that my previous mentors at that time - one being a veteran game developer who worked on Assassin’s Creed and the Hitman franchise and the other one from big tech - did not encourage me to join big tech, but rather to build things that I really believed in and I’m passionate about.
How important do you think mentors are? How do you find them?
Mentors are extremely important because they've already been in your shoes and have asked the same questions you're asking yourself now. You might eventually get to the same point without a mentor, but with one you can get there faster and advance more rapidly.
In my case, many mentors actually found me. I would go to different events, say hi to people, engage with them. I was young, but I would talk passionately about all the things I was building and was always asking questions. This impressed people and some of them offered to mentor me.
I think everyone should have the courage to proactively reach out to people and ask "Hey, do you want to be my mentor?"
If you want to have a mentor, be very clear about what you'd like to learn from them. Choose someone who has a career path or achievements that you aspire to. Then prove to them that you're worthy of their time investment.
Internship experiences at Google
So your first Google internship was YouTube in Zurich. How did it go?
It was funny because in the first year of university, you don't know much about networks, servers, and a bunch of related concepts. Yet there I was, tasked with developing a MapReduce pipeline for data migration - which was not easy.
There were many things I didn't know because no one had taught me before at university. It started a bit rough, but got exciting and I was learning a ton - there were so many brilliant people at Google to learn from and aspire to be like.
What made it unique, however, was access to Google’s monolithic codebase, seeing all the incredible things people had built before me. Google has one huge, not restricted, repository, and it was fascinating to just sit for hours and read through what other people had done.
I could look at design docs from almost anyone. It might be more restrictive now, but at the time I could easily search through and learn about the stack of YouTube, Google Shopping, Ads, etc., their approach to problems and the solutions they designed.
What's your advice for students trying to get into big tech and maybe even interviewing at the moment?
I was lucky because I had done competitive programming for many years, so the interview questions were very similar to problems I would usually solve. If you grind Leetcode for many months and do it rigorously, you should be able to pass the technical part of these interviews.
Then, it becomes about practicing, doing mock interviews, to come out of your shell and be more outspoken about your thinking process during the interview. This is actually more important because the interviewer is really evaluating if they want to work with you as a person.
If you're very anxious in the interview or too shy to speak up, then it's not going to come across well. Usually companies are looking for interns who are joyful to work with and bring positive energy to the office.
You learned many specific things like MapReduce during your internships. Are these skills valuable later on or are they hyper-specific to that use-case?
I was a bit worried about this because I did 5 internships at Google and Google is somewhat of a technology island. Most of the stack is built specifically for internal use at Google and not really used anywhere else (besides a few open-source projects).
Looking back, my concerns that the knowledge I gained was not transferable were not really true. The engineering principles and the technical stacks seemed to be quite similar across and surprisingly universal. When learning or working on something new, I ramp up very quick given that experience - what one of my mentors used to describe as “reading code like poetry”.
If someone wants to build their own start-up, is there still something valuable that they can learn from a big tech internship?
If you want to build your own startup, you need a lot of skills that are hard to gain as a SWE in big tech. Particularly, as an engineer within a big engineering org, you rarely get to interact with product or marketing people before becoming very senior.
In fact, you are usually working on a fairly isolated project that is high-impact: the numbers are massive, there was proper planning before and the engineering challenges are fascinating. However, you don't get much insight into broader company strategy, roadmaps, or the high-level decisions being made above you that explain where what you’re building fits.
Understanding these areas is very important for building your own startup, but you don't really get exposure to them in most big tech internships. Overall, you learn solid engineering skills that can take you a long way, but miss out on a important, yet less technical learnings.
You interned at Google Los Angeles and New York. A big question for many European students - how did you land an internship in the US?
It might be a little harder due to competition, but honestly not that much. The main difference is that the application timeline is quite different. To apply for internships in the US, you usually need to apply by mid-September. For Europe, you typically have until late October or even November for the big tech companies.
Regarding immigration, US internships require a J-1 visa, but major tech companies are able to sponsor without difficulty, as it is temporary. It's way harder to get a full-time job though due to the H-1B visa’s lottery system and even large tech companies cannot guarantee that.
Why did you want to go to the US for an internship? What's the difference between working in the US and Europe?
I had two main reasons for heading to the US. First, to experience living there - pretty standard bucket list stuff. Second, due to the assumption that people in the US tend to be more business-minded, and I wanted to learn from that environment.
This assumption was true to some extent. In Europe, I found we’re often deeply focused on technical excellence. You have brilliant engineers building awesome stuff, but there’s way less discussions about the broader impact and the go-to-market of what they're building.
The US offices feel different because you are commonly surrounded by directors, VPs and other high-level stakeholders so these business conversations happen all the time. This naturally leads to more impactful projects that get good support from the high-ups, since the decision-makers are all in one place.
There are a lot of crucial conversations and decisions that happen in the office over casual lunch breaks. If you're in Europe and your VP is in the US, you miss out on those informal chats that help you understand the thinking behind decisions and last-minute chats.
Why did you stay at Google for five internships? Why not explore opportunities at other companies?
The ability to be very flexible in choosing my projects was extremely appealing. My first internship was fairly generic, but I was also very young at that time.
After that though, I got to work on ML orchestration, ML performance, NLP, deployed ML in production, and even research. Going into a next internship, I had a great amount of freedom to choose what I wanted to focus on. It's rare to have that level of choice in a first internship at a company, whilst for returning interns, if you put in the effort to network with a lot of teams and knock on a lot of doors, you can find the ideal fit for your interests. It doesn't come for free though. You have to put yourself out there and set up a ton of calls.
Masters at EPFL
You came to EPFL for your masters. You could have stayed in Romania or even gone to the US, UK. So why EPFL?
The cost of education was a huge factor, so I didn’t consider US or the UK. The tuition is relatively affordable in Switzerland and Germany.
I didn’t consider ETH because they don't have a summer break, their exams run in late August, thus hard for internships. Therefore, I found EPFL to be the best option and it also seemed better than TU Munich.
Also, I wasn't ready to start ‘adulting’ yet. After the bachelors, you basically only have the choice to start a full-time job and I thought you’d learn much less then. Doing a masters was a way for me to extend my student years and have a few more internships to explore different things. I really wanted to get some exposure to research as well.
At EPFL, you joined organisations like Founderful Campus and EPFL-UNIL Entrepreneur Club. You already had an impressive experiences at Google, so why did you join these organizations?
I'm always seeking to maximise my learnings - as I’m very inquisitive by nature and these groups seemed super valuable, in terms of learning opportunity and the people I could meet. It's the same as before, a group of super passionate people coming together to build something meaningful.
The students I met were brilliant and really fun to work with. When I joined the EPFL-UNIL Entrepreneur Club as a startup manager, there was a lot of room to build new programs and infrastructure. The startup ecosystem at EPFL was quite limited at the time - initiatives like Blaze didn't exist yet. So we felt we were doing something impactful by gathering people and enabling them to collaborate on projects.
Later, I joined Founderful Campus as a managing partner, which was an incredible experience. I initially doubted if it made sense for me, since I've always been more passionate about building things rather than the venture capital side. However, it turned out to be hugely eye-opening and gaining exposure to the VC mindset is a critical asset for a founder - it teaches you to think more strategically and how to fundraise.
Especially, it forces you to think not just about what you're building for the next 2 months, but the long-term vision, where you want to end up, and how to actually get there. This perspective is immensely valuable for founders.
Being part of Founderful also sets a very high bar for any startup you may start or join in the future. You get exposed to so many founders and develop a taste for what makes a great founder and founders you’d aspire to be like. I recommend joining Founderful or other campus funds to all aspiring founders!
Working on Research
One of the main reasons you came to EPFL for research. What do you think about research after experiencing it?
Research is definitely exciting and the freedom of choosing yourself which questions to pursue is very empowering and intriguing. However, I love a little bit more to build things that get into the hands of users very quickly. It’s the small dopamine hits when you launch something new and see immediate user impact - those moments that confirm that you’re moving in the right direction.
Research requires patience - it is more of a long-game where you need sometimes years of dedicated work before seeing tangible impact. If you're fortunate with your lab, your professor will provide great direction and guidance to help you reach that point. But it can be hard to find.
For me, that pace and somewhat the solitary nature of research made me go back to building. I love the autonomy to build, launch, and iterate rapidly while getting feedback from real users. That tight feedback loop brings me a lot of fulfilment. In research, you also need to run a lot of experiments, see them fail and still continue until eventually something works - but I agree that it feels unearthly good when it does.
You did your master's thesis at Google DeepMind, how was that experience?
It was an incredible experience that I am very grateful for. I joined a research team focused on explainability, data selection & attribution as well as active learning. I was really passionate about explanability and NLP and that's what motivated me to start emailing a lot of researchers at Google working on similar topics.
It is true that some did not reply and some gave me very direct response, saying there was no reason to take me when they could get an more experienced PhD to fill the same headcount. But luckily, I was found a perfect fit for my research interests.
Besides the great mentorship from experienced researchers that were very hands-on, you also have access to a lot compute resources. In contrast, some of my friends doing their thesis in other places had challenges training models because they had to share resources and things were moving painfully slow. That is an unfair advantage in research in general.
At Google, the only limitation was how quickly I could code and read papers. So the experience was hyperproductive. During my internship, ChatGPT was released and it was also fascinating to see how an organization as large as Google could rapidly pivot and start catching up.
Joining OpenMined
Why did you decide to join OpenMined instead of continuing at Google?
There are a lot of talent and brilliant people you can learn from at Google and I benefited from it greatly. However, as with many large organisations, things tends to be slow-moving and bureaucratic. For myself, I’ve found I thrive in environments that move quickly, where I can assume a high degree of autonomy and I can solve problems holistically.
Moreover, Google was solving a different set of problems than the ones OpenMined is addressing. The focus on critical challenges around data access, information governance and enabling broad listening (as opposed to broadcasting information paradigm of web2) were compelling to me. Moreover, I have deeply inspired by the people I am working with now everyday, so it was an easy decision to move forward with OpenMined.
How did you find OpenMined and what do they do?
I've actually been an open-source contributor to OpenMined for about 4 years now. During my time at EPFL, I would periodically submit pull requests and contribute code, docs. What I found there was a small group of people who are extremely passionate about what they are doing.
There's no for-profit structure — everyone is in it because they believe in it and it's an awesome feeling to be in such an atmosphere everyday. It’s less of a "9 to 5" schedule and closer to a hectic startup environment sometimes - as we’re working against a quite ambitious mission.
Essentially, there’s a lot of questions we could answer, as humanity, today, but we cannot because data is siloed across a lot of institutions, companies and individuals, due to privacy risks or simply because they are valuable IP. If you think of machine learning, access to data (again!) and not necessarily compute is the incoming bottleneck for frontier models. Breaking such data silos could help us advance breast cancer, understand how watching Netflix impacts our health or solve much harder socio-economic problems. However, that needs to be done in a responsible way - without walled gardens and with appropriate governance.
In a nutshell, OpenMined builds the technology that could make that happen. If we get this right, there might be 1000x more data in every scientific field - that any researcher, regardless what university or company they work for, could use in their work. We could make so much more progress on really important problems. So that's what drives us.
Advice for students
What do you think about students dropping out to found a startup or join one?
University is awesome because it gives you time just to learn. Whereas when you start working, especially towards something specific like a start-up, the pace is very, very fast and the luxury to stop for a moment and study something that interests you is gone. There’s a time for learning and I believe another one for doing - having solid foundations is important.
Looking back, do you have any advice you'd give to your younger self?
There’s many things I would tell my younger self, but the most important is to trust your abilities more. Lack of confidence can be paralysing - it’s hard to share your work and speak up about your achievements. You end up always second-guessing, which fuels you - but also takes away some courage to do more ambitious things.
A good way to build that confidence is to share your work and get feedback. For example, in high school I built with a few friends a VR version of Hearthstone inspired by Yu-Gi-Oh. Only the game design ended up being 100 pages and the visuals were really cool. However, perfectionism held me back - I never released any of it because it didn't feel "ready". I guess sharing small achievement is great for growth and learning.
Do you feel like you have your future path all figured out now?
Each of us goes through an exploration phase and it takes time to reach a point when you know exactly how you want to spend the rest of your working years. I certainly don't feel like I'm at that point yet.
I don't know exactly what I'll be working on next. But I do ask myself every week - “given my current skills and the time I'm willing to invest, what could I build or achieve that would make me feel really proud?” So that when I’m 90 years old and looking back, I’d be very satisfied with how I spent my time and all the little steps I took.
Basically, whenever I come across something that I think would give me a higher sense of accomplishment and fulfilment, I jump on it. So far, each experience taught me a ton of new things and abilities. Over time, these compound and you're able to take on bigger and more impactful challenges. I have trust in that process and let it guide my path.
Hey - Arnie here!
Thanks for reading this week’s month’s year’s (lol) edition of the blog!
As you can tell with the inconsistent posting - I’m having a hard time keeping up with writing blogs & all the other projects going on.
If you’re very passionate about the work done here and would like to help out - send me a message on LinkedIn or email: anirudhh.ramesh[AT]gmail.com!
In the message, tell me your biggest achievement and why you’d be a great fit to help run the blog :)
What you can learn:
Mentorship, guidance, introductions from me (Arnie!)
Opportunity to speak with excellent alumni from EPFL & ETHZ, to jumpstart your career & ambition!
Requirements:
Super passionate about the blogs!! You should have read basically all of them :)
Early bachelors/masters/PhD student (at least 1 until graduation)
Excellent English-writing skills
Also, as always: