Three ways of making AI more dependable.
I study how to make modern AI systems safer and more reliable - especially when they interact with people, tools, and the physical world. The work combines rigorous evaluation with practical safeguards.
Enter the research atlas →