소개
I’m a Sr. Principal Research Manager in the Deep Learning Group at Microsoft Research, where I lead a team focused on AI modeling and agents to improve human–AI collaboration. Our work bridges AI and HCI, spanning LLM post-training for human–AI interaction, agentic knowledge discovery, agentic interfaces, user simulation, adaptive UX, and AI for education.
Recent work has been recognized with Outstanding Paper Awards at ICML 2025 (CollabLLM (opens in new tab), an LLM trained for proactive collaboration) and NAACL 2024 (opens in new tab) (on self-improving conversational LLMs). Earlier contributions include co-developing one of the earliest neural conversation models (opens in new tab) (2015), pioneering work on retrieval-augmented dialogue (opens in new tab) (2017), some of the earliest widely used open-weight conversational LLMs (DialoGPT (opens in new tab), 2019, GODEL (opens in new tab), 2022), and the MathVista (opens in new tab) multimodal reasoning benchmark. More information is available on Google Scholar, (opens in new tab) LinkedIn (opens in new tab), and in my CV (opens in new tab).
Academic Background:
2007-2009 Postdoc in Computer Science, Stanford University, Supervisor: Christopher Manning
2007 Ph.D. in Computer Science, Columbia University, Dissertation advisor: Kathleen McKeown
2002 M.Sc. in Computer Science, Columbia University
2001 Diploma in Computer Science, EPFL