University of California, Berkeley
UC Berkeley is one of the leading universities for computer science and machine learning in the United States. My work there was more research-oriented and covered several parts of modern AI: agents and multi-agent systems, robotics and human-machine interaction, multimodal learning, biomedical AI and model evaluation.
Across this period, I worked on a diverse set of problems: how models represent complex evidence, how they behave when information is partial or noisy and how their behavior can be evaluated with enough scientific care.
Those questions connect directly to my broader research interests across agent workflows, robotics and human-machine interaction, multimodal evidence, biomedical AI and evaluation methods for systems whose behavior unfolds over time.
Over time, my Berkeley research moved through several related areas: agents, robotics, multimodal learning, biomedical AI and model evaluation.
Academic Work
Core areas
- AI agents and multi-agent systems
- Robotics and human-machine interaction
- Biomedical AI and computational pathology
- Transformers, LLMs and multimodal learning
- Model evaluation and reliable machine learning
Academic work
- At Berkeley, my academic work became more research-centered and moved across AI agents, robotics and human-machine interaction, multimodal learning, biomedical AI and model evaluation.
- I studied AI systems that reason, interact, coordinate, learn from multimodal evidence and adapt when information is incomplete.
- I kept returning to common questions about representation, robustness, evaluation and useful behavior under uncertainty across agentic AI, robotics, biomedical AI and multimodal learning.