By HANNAH MARTIN
Mona, a Student Assistant at idalab, explains why AI’s habit of turning small discoveries into game-changing applications drew her to Data Science
Hi Mona, thanks for letting us ask you some questions.
That’s okay. Thanks for taking an interest!
So let’s start. How did you end up working at idalab?
I saw the idalab opportunity on a jobs portal, and liked the idea of seeing Data Science in practice along with Strategy and Biology. I wanted to see how AI could be applied to solve real-world problems. From the website, the working environment looked creative and I liked the different educational backgrounds of the idalab team. My colleagues have studied Biology, Maths, Design and even Film.
As for me, I did a Bachelor’s in Economics. And now I’m doing a Master’s in Statistics. There’s an ongoing trend for people studying these subjects to go on to work in Data Science. It’s trendy because it’s often at the cutting edge of research and you can see the impact of your work quickly. The speed of advancements in Data Science means that research papers from just two years ago are often already out of date!
Applying data science and seeing AI in practice takes many people towards working for one of the large consultancy firms. Did this tempt you?
I hear stories from classmates about what it’s like. Working for a big company — with lots of people, complex projects and moving parts — can often be very task-orientated. You’re given small task after small task, but you never quite see the bigger picture.
I believed that in a smaller company I’d get more opportunities to learn. And so far this has been very true. I’ve been given my own projects and lots of responsibility. Of course, I get help along the way. So if I’m stuck, I know who to ask. Even though idalab is a small company, it has a lot of clear structures in place to help you succeed, especially when you might not have a lot of working experience.
idalab has a focus on Life Science. Was there anything about Life Science that attracted you?
Yes, working in Life Science, there’s the potential to see the influence of AI on some of humanity’s biggest problems – drug discovery, climate change and more. Data Science discoveries can have a domino effect, starting a chain of events that eventually has a big impact on the world. For example, we recently saw a huge discovery where an AI tool developed by Google DeepMind solved a 50-year-old problem about protein folding.
You’re participating in the Junior Associate Programme, on the Data Science track. Could you briefly describe what this programme is all about?
It’s a programme aimed at graduate students. The idea is that you start as a Student Trainee like I am now. And then there are four tracks to choose from: Life Science, Data Science, Engineering and Strategy. It can be full-time or part-time. I work two days a week, which I can arrange however fits best to my university schedule.
One of the structures we have in place to help us is idalab’s mentorship programme. My mentor is Benjamin. He’s my go-to for any questions. We meet every three weeks, discussing several points — how I’m balancing my work and university life, ongoing projects and what I’m learning, for example.
What projects are you currently working on?
I’m near the end of a project with a multinational Pharma company. It’s a large project with a lot of people involved and a lot of different tasks. By combining information from a medical database with a commercial database, we’ve developed a tool to help them identify biotech companies doing pioneering research. It can identify emerging trends in drug discovery, therapies and biomedicine.
And what’s your role specifically on the project?
I’ve taken on many different roles so far at idalab and enjoyed the variety. On this specific project, I was a programmer, so I was involved primarily with databases. As I mentioned, we had these two sets of data: commercial data and biomedical data. I created a Machine Learning model to link the two together. I was used to working on smaller projects with CSV files, but this was on a much bigger scale. I used SQL, Python and editors such as Visual Studio Code and Datagrip that I was introduced to by my colleagues; they are way more effective than the ones I use at university!
I also worked with my colleague and the project’s Lead Data Scientist on analysing the data, discussing the project together and running test scenarios on our algorithm. I was in the client meetings from the very beginning. This helped me practise interacting with people from different organisations and presenting my work.
On the idalab website, it says: ‘Uncertainty and complexity is the nature of our work and strategy is the answer.’ So what does this mean, exactly?
For me, the best way to explain this is in reference to the project I was just talking about. Our goal was to detect trends, but what is a trend? A trend has to be defined. And the strategy is how we create this definition — how we remove the uncertainty. It’s only then that a trend can become a curve on a graph or the segment of a chart. In our project, using the two data samples, we developed two indices to spot the trends and sound an alarm when one was detected.
How do you learn new things at idalab?
idalab has an emphasis on learning and development, with lots of training, seminars and a three-day retreat twice a year. This helps ensure everyone is on the same level and helps you acquire skills you might not have picked up at university. In December, I started with methodology training, which involves studying evaluation metrics that aren’t so common.
Your development is then monitored using an incredibly detailed competence framework, so it’s easy to identify where you still have the potential to learn. Once you’ve reached a certain point, it’s possible to be promoted to a Junior Associate.
In the last year and a half, I’ve also learned a lot from my coworkers. My colleague, Gillian, for example, who was the project lead on my last project team, helped me to get a feeling for our client’s needs and how we should structure our approach. Then there’s Rouven, who introduced me to technical topics and conducted code reviews with me.
You mentioned idalab’s retreat. Can you tell us more?
We usually go to the middle of nowhere. Here, we’ll discuss organisational development and broaden our horizons on topics such as different business models. There will normally be some technical training too – plus Frühsport, which I try to avoid.
You said the retreat was in the middle of nowhere. Are we talking about camping here? Were you foraging for food?
Luckily, it’s not camping! Although the last retreat was virtual, because of Covid-19, it’s usually at a beautiful mansion estate in Brandenburg with a big garden, a barbecue, conference rooms and plenty of space for everyone.
Last question: could you grant us a glimpse of your calendar?
Sure! Let me talk you through this typical week.
I currently work on Mondays and Tuesdays, with my studies taking place on the other days. On Mondays, we have our weekly idalab meeting with the entire team. We go through people’s news, cover any questions and, if needed, enlist help from elsewhere in the team to solve any challenges.
Later in the day, we have a ‘Fix’ for the project I’m involved in. After that, I meet up with Luis to discuss the AI x Life Science Forum. The Forum is a monthly event hosted by idalab where a team member presents a research paper showcasing AI in medicine. It’s a great way to stay up to date with current research. Luis and I are currently working on a process that ensures we never miss a great paper!
In the afternoon, I have a one-to-one with my mentor, Benjamin. For the rest of the day, I’ll work on the results of an internal review workshop that we conducted. The review relates to a project that’s nearing completion, and we’ll present the results to the team on Friday.
Of course, there’s lunch with colleagues in between, on our big dining room table. I’ll spend some time on methodology training about evaluation metrics. On Friday, I’ll discuss what I learned with Paul. And finally, we’ll end the week by talking through any lessons learned from the project review.
Mona is studying for a Master’s degree in statistics at Humboldt University, Berlin. She first completed a Bachelor’s degree in Applied Mathematics and Economics at ENSAE ParisTech. Throughout her studies, she’s demonstrated how AI and Machine Learning can be used as a force for good — from creating poverty prediction models to detecting Islamaphobia in tweets.
Now, Mona’s bringing this aptitude to Life Science. She’s part of our Junior Associate Programme, where, despite being on idalab’s Data Science track, Mona’s work includes much more than coding. You’ll usually find Mona preparing presentations for clients, working with large datasets and collaborating with different team members across Life Science and Strategy.