PhD student Lorena about her experiences in the AI4FinTech Research Lab

PhD student Lorena about her experiences in the AI4FinTech Research Lab

May 2, 2024
Written by

For our first ICAI the Spotlight blog post, we asked Lorena Poenaru-Olaru, fourth-year PhD candidate at TU Delft's AI4FinTech research Lab (AFR), to share her academic journey and work experience within the AI4FinTech Lab.

Lorena recalls a key moment during her Master's thesis collaboration, realizing how blending academic methods with industry knowledge can make a big difference. This realization inspired her to seek a PhD position that combines research with practical business needs:

“A few years ago, I was working on my MSc thesis in AI in collaboration with a company. I began my project by applying state-of-the-art academic techniques to industry data and noticed that they were not performing as well as expected. After long discussions with my colleagues who were domain experts, I realized that these techniques were missing some domain context. After combining them with context information, the performance of these techniques increased significantly.

At that point, I realized the enormous value of combining academic research with domain expert knowledge and this motivated me to search for PhD positions in similar settings. I found out about the open PhD position in the AFR lab and thought “This is exactly what I am looking for!”. 

“When I started my PhD, I wanted to research a problem that AI practitioners are constantly facing. I began by discussing with multiple AI practitioners at ING about their current challenges when deploying AI models in production. Most practitioners were concerned about the reliability of their AI models over time since their performance is significantly impacted by data changes, a common phenomenon in practice, and there was little guidance on how to monitor them. While reading scientific articles, I noticed that different data monitoring tools were developed on synthetic and open-source datasets, but there was no study to assess whether they apply to the industry. At that moment, I noticed that there was a gap between industry and academia in building reliable AI systems that I could bridge with my PhD research by guiding practitioners into which data monitoring techniques to employ and how to assess their effectiveness on their data.”

The most rewarding part of Lorena's journey is seeing her research directly help businesses. When industry experts confirm that her findings are useful, it proves her approach works and can improve AI systems used in business. In addition to academic learning, Lorena stresses the importance of picking up practical business skills during her PhD. From managing projects effectively to collaborating well with teams, she believes these skills are crucial for success in any field.

“The experience of pursuing a PhD is useful to earn valuable transferable skills, such as project and time management, critical thinking, and research skills. Although sometimes, PhDs collaborate with other academic researchers, and most of the time the PhD is seen as a “lonely journey”. 

“One of the most valuable experiences was collaborating with a globally distributed team (Poland and The Netherlands) and discovering so much about effective communication across different nationalities. Doing a PhD in an ICAI lab allowed me to develop my ability to work in a team.”

Lorena invites everyone to explore the impactful work happening at AFR and its potential to make a difference in the world of AI and FinTech.

About the Spotlight

Each month, ICAI shines a spotlight on one of its labs, offering insights into their research pursuits, accomplishments, and the individuals driving innovation within these labs. Additionally, every month, ICAI hosts a meetup event at the location of the featured lab, providing attendees with an opportunity to delve deeper into the lab's activities in an engaging real-life setting.


Stay in the loop

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.