Fostering AI architects: idalab’s first Data Strategy Summer Fellow
While Artificial Intelligence is continuing to transform the world as we know it, the need for “AI generalists”, who take the role of architects designing custom solutions becomes ever more acute. It is no wonder that AI architects are in short supply: AI architects combine profound expertise in AI-methodologies with a highly analytical, yet creative and solution-focused mindset, enabling them to see the bigger picture and make strategic decisions. Today, only few people have this kind of generalist skill set.
We believe that there is great talent in academic fields outside of Computer Science, Mathematics and Statistics. Providing development opportunities for the next generation of AI leaders is one of our top priorities. A few weeks ago, Oxford graduate Lisa Martin, an economist by training, joined our team for a two month fellowship as Junior Data Strategist. Kirstin Taufertshöfer is a Data Scientist and Lisa’s mentor at idalab. She spoke to her about her ambitions, perspectives and projects.
Kirstin: Lisa, why is the world of AI and Data Science appealing to you, your background being in economics?
Lisa: At different stages of my academic and early professional career, I realized how data science impacts and transforms research, businesses and policy making. While I’m of course not a data scientist, I’ve always taken an interest in the field. Within economics, my focus has always been on statistics and econometrics – at the University of Tuebingen and later in Nuremberg and Oxford. During the course of my studies, I experienced first-hand some of the versatile use cases for predictive analytics. Let me give you two examples from finance and development economics: For a seminar thesis at Nuremberg University I did a project with Consorsbank, predicting trading behaviour using macroeconomic and stock market indicators. In Oxford, I used synthetic control methods in my Master thesis about ethnic inequality and civil war. I’ve always enjoyed empirical research, using “real world data” and statistical methods. I’m intrigued by how data science can enhance our problem-solving and decision-making capabilities. As a student assistant and business intern I had already experienced work in academia, the automotive sector and management consulting. When I graduated from Oxford this summer, I was looking for an opportunity to explore what kind of impact someone with my background could have when working in the field of Data Science.
Kirstin: And why did you then choose to approach idalab?
Lisa: For one thing, I thought it was exciting that the team members’ backgrounds are so diverse. They range from Computer Science, Maths and Natural Sciences to Linguistics, Theatre Studies and Design. idalab is of course specialized in Data Science, but they generally take a generalist approach, seeing the bigger picture and thinking out of the box. They are architecting the best solutions for clients, carefully considering each AI use case individually. They don’t just “throw” AI at every problem they come across, simply because they could. You can easily get the feeling that that is what most AI initiatives do – and that is why, so often, they’re failing!
Another thing I like about idalab is their ambition to participate in and shape the public discourse about AI. For example, they regularly arrange open seminars to discuss the work of scholars and thought leaders in Data Science. And of course, Berlin is a great city to spend some time in. It’s a vibrant AI hub that also has a lot to offer culturally.
Kirstin: What exactly will you be doing during your time at idalab?
Lisa: When I started, I sat down with our manager Paul and discussed how we could structure my time here so that it would be meaningfully spent for both idalab and myself. Paul proposed that I should spend time on thoroughly researching a topic that is growing to be very important for idalab and their clients: ways of preserving data privacy when applying AI methodologies. I’m now researching this topic, which involves independently conducting expert interviews with executives and start-ups. This way I’m expanding my network and meeting a lot of very interesting people.
In addition to working on my own project, I’m learning a lot about the projects and industries my colleagues are involved in, and about the cutting-edge AI techniques they are using. I also regularly get to participate in calls and will be joining my colleagues for meetings with clients.
Kirstin: Could you elaborate a little on the topic you’re researching for idalab and why it is so important?
Lisa: In the light of the new legal framework GDPR, data privacy has become one of the most pressing issues in data science. I’m researching strategies and technologies that enable data scientists to protect people’s privacy while continuing to gain valuable insights from the data. These approaches include improvements on existing technologies in anonymization and pseudonymization as well as exciting technological advances such as fully homomorphic encryption – a way to facilitate computation on encrypted data so that sensitive data never has to be “out in the open”.
Kirstin: What are the central goals of this research?
Lisa: Two things, really: developing expertise and establishing contacts. Right now, only very few people have a good overview of the relevant strategies and technologies. As an agency for data science, idalab needs to stay on top of this issue, being able to provide their clients with solutions when needed, for example by bringing them together with software providers and experts in the field. As I mentioned, as part of this research, I’m conducting a number of interviews and benchmarking different solutions so that idalab will be perfectly positioned to advise their clients on any data privacy related issues.
Half-way through my time at idalab I will be presenting my preliminary results to all my colleagues and get their feedback. I will then be preparing a white paper about my findings, which I’m also thrilled about. To me, this proves that my project is of real importance to idalab, since it has the potential to positively impact idalab’s work and our clients.
Kirstin: Lisa, thank you for your time and have a great two months in our team!
Lisa Martin recently graduated from the University of Oxford with distinction, having completed the MSc in Economics for Development. Prior to this, Lisa studied economics with a strong focus on econometrics in Tuebingen, Boston and Nuremberg, completing a BSc in International Economics within the top 5% of her class. Lisa has worked as an intern in the Economic & Market Intelligence Department at Daimler in Stuttgart, as well as in management consulting at Oliver Wyman in Munich. For her studies, she has received scholarships from the German National Academic Foundation and the German Academic Exchange Service. During her time in Tuebingen, Lisa was president of oikos Tuebingen e.V., a student organisation for the furthering of sustainability in economics.