What attracted Dr. Ana-Maria Cretu to CISPA was the unique opportunity to work alongside so many people in security and privacy. “There are many opportunities for collaboration, but also for learning from others simply by being in such close proximity to them,” Cretu says. “I think that’s a great environment to be in.” She was also impressed by the conversations during the application process and the atmosphere they conveyed: “I was pleasantly surprised that despite CISPA’s size, I never had the feeling that it was overly competitive. Instead, everyone seemed genuinely helpful. That was one of the factors that convinced me to join.” She was not only thinking about herself, but also about the environment she could offer her future doctoral researchers. “Besides the generous salary, being part of CISPA means they would learn much more than just their own research topic. They would be exposed to many different security and privacy topics, which is something I had less of during my own PhD because there are fewer researchers working on these topics at the university compared to a topic-focused center like CISPA.”
Looking Behind the Façade of General-Purpose Solutions
At the heart of Cretu’s research is a simple but fundamental question: Can increasingly powerful AI systems deliver on their promises without compromising privacy, security, and safety? She is particularly interested in technologies that promise general-purpose solutions, such as text-to-image models. “The broader question that interests me is whether we can build systems with these general-purpose capabilities without sacrificing the properties we care about, such as privacy, security, and safety,” she explains. Rather than taking the capabilities of modern AI at face value, Cretu advocates for a more rigorous mindset. “What I care about is taking a more principled approach—carefully evaluating systems before they are deployed instead of simply assuming they will work as advertised.” Ultimately, she believes that thorough evaluation is the key to building AI systems that are both capable and trustworthy.
New Ideas for Her Research Group
With the start of her tenure-track Faculty position at CISPA, Cretu is now building her own research group. Many of the themes she has explored over the past years will continue to shape her work. One of her priorities is to deepen the scientific understanding of synthetic data. “In the area of data privacy, I think synthetic data is one of the most promising research directions,” says Cretu. “It has become a very popular technology in recent years, particularly for sharing data in a GDPR-compliant way. One of my goals is to contribute to a more rigorous scientific understanding of synthetic data.” Another area the group will work on will be AI safety, particularly the safety of generative image models. “I would like to continue the work I have already started in that field and develop methods for evaluating and improving these systems,” she explains. She also wants to collaborate closely with organizations that have clearly defined privacy and security requirements. “Working on concrete use cases allows research to have a direct impact outside academia,” Cretu is convinced.
From Mathematics to Applied Fundamental Research
At school, Cretu developed a passion for mathematics, taking part in competitions with great success. “I started participating in the Mathematics Olympiad when I was twelve. I finished second in my county and from then on I continued participating. I’ve always loved mathematics and in high school my dream was to become a mathematician.” Studying mathematics was therefore the obvious choice, especially after she received a scholarship to study in France. “It was really my dream because France is one of the best places in the world for it,” she recounts. “At some point, though, I realized I wanted to work on something more applied,” Cretu continues. “I couldn’t really see myself spending my career working only on abstract mathematics. Before starting my master’s degree, I gradually moved toward computer science, particularly machine learning, and I’ve stayed on that path ever since.” Even so, mathematics continues to shape the way she thinks and conducts research today. “My background in mathematics has given me an interest in the formal treatment of security, privacy, and safety in the context of AI alongside empirical evaluation, and in building an understanding of why vulnerabilities exist and whether they are inherent in a system.
Cretu’s Approach: Fewer Projects, Greater Depth
“At the moment, academia often feels like a numbers game. There is constant pressure to publish more papers and submit to more conferences,” Cretu says. “My own philosophy, which I learned from my mentors, is different. I think it’s better to do fewer projects, but to pursue them in greater depth and with greater ambition.” This is also the philosophy she hopes to pass on to her students. “I would like my students to feel that they have the freedom to focus on work they are genuinely proud of, even if it takes longer. I’d rather see two or three substantial contributions than several incremental papers.” For Cretu, this philosophy goes hand in hand with a clear understanding of what good science should look like: “Research should be rigorous. It takes time to ask the right questions, evaluate ideas carefully, and avoid cutting corners. Those are the values I would like my lab to embrace. In a way, it’s about getting back to the basics of good science: taking the time to do things properly and being proud of the work we produce.”
Career Opportunities in Europe
For Cretu, accepting the tenure-track position at CISPA also means having the opportunity to continue her career in Europe while staying close to her family in Romania. She also appreciates the proximity to France, where she completed her bachelor’s and master’s degrees thanks to scholarships. The only thing she regrets is that French is not spoken more in Saarbrücken. She appreciates the free German language courses offered at CISPA all the more. Outside of research, Cretu enjoys bouldering. But her greatest enthusiasm remains the work itself. “I chose an academic career because I genuinely enjoy research and I do not see myself doing anything else.”
