Dr Tereza Iofciu gives insight into teaching AI technology

30 November 2021

SERIES (5): "Female AI - four questions to women in AI" - interview with Dr Tereza Iofciu, Head Coach Data Science at Neue Fische

Artificial intelligence (AI) offers all kinds of exciting career paths, especially for women. The Hamburg-based AI company Synergeticon, the ARIC Hamburg association and the proTechnicale educational initiative have joined forces to promote women in the AI industry and to raise their visibility. Women who have already gained a foothold in the industry share their experiences in the Hamburg News interview series "Female AI - Four Questions to Women in AI".

Dr Tereza Iofciu is Head Coach of Data Science at Neue Fische and speaks regularly at global AI conferences. She founded the PyLadies Hamburg group, organises numerous meetings and is a role model for the participants of proTechnicale, the technical gap year for young women in Hamburg. Iofciu was also an ambassador for Women in AI in Hamburg. Before joining Neue Fische, she worked as a Data Scientist and Data Engineer at FreeNow (formerly My Taxi) and Xing SE (now New Work SE). In 2011, she completed her PhD in computer science on information retrieval at the L3S Research Institute in Hanover. In her spare time, she draws dinosaurs called tiyepyep.

Hamburg News: What do you do in the AI industry and why is it important?

Dr. Tereza Iofciu: As a Data Science Head Coach at Neue Fische, I teach AI and machine learning to different groups of people. Each bootcamp lasts three months. The participants are either entering the field of data science from academia or other industries or want to improve their knowledge. An important aspect for me is that participants leave the bootcamps with an understanding of the responsibility that comes with data science knowledge. Machine learning and AI are powerful tools that can solve many large-scale problems, but can just as easily amplify problems.

I also volunteer in various organisations, I am a joint organiser of PyLadies Hamburg, vice-chair of the Python Software Association and part of the Diversity & Inclusion working group of the Python Software Foundation. Python is one of the most popular programming languages used in Data Science.

Hamburg News: How did you get involved in AI?

Dr Tereza Iofciu: I was in the right place at the right time. I was good at maths and physics in school, so I decided to study computer science in Bucharest, Romania. It seemed like a good thing for the future in 2000. At the end of my studies, I had an opportunity to do my thesis at the L3S research centre in Hanover, where I also completed my PhD on Entity Recognition, Recommender Systems, User Matching. The field of Data Science emerged at the time.

Hamburg News: What are your (professional) plans for the future?

Dr Tereza Iofciu: I like learning, teaching and increasing the diversity of qualified Data Scientists. My role at Neue Fische is very much aligned with these values. On a personal level, I have started learning Data Driven Documents, which combines data and graphical elements (D3.js) and improving my data visualisation skills. I feel that the general population is more exposed to data visualisation than ever amid the pandemic and there is plenty of potential to improve data literacy in the world.

Hamburg News: Why should there be more women in AI?

Dr Tereza Iofciu: We can all agree that the general population is far more diverse than teams that build the AI systems with which we interact. In 2018, women accounted for up to 16.58 per cent of tech professionals in Germany, with a wage gap of 25 per cent (Honeypotio's 2018 Women in Tech Index study). In most of my past jobs in Germany, I was often one of the few women on the team. The farther up I climbed the career ladder, the worse it became. There are many stories of AI systems not working well for people from underrepresented groups, people of colour, people with disabilities or women.

On the other hand, data products created by diverse teams have a better chance of appealing to a diverse population. It's pretty easy to transfer our personal biases into AI systems, and having different perspectives represented on the team is a crucial safeguard against inadvertently biased systems. If that's not reason enough, new research shows clearly that companies with a diverse workforce perform better financially (McKinsey's 2020 study "Why diversity matters").
ae/tn/sb