À propos
I’m a Principal Applied Science Manager on Microsoft’s CoreAI team, where I lead an applied research group dedicated to boosting developer productivity with ML et AI. We’re currently advancing GitHub Copilot’s capabilities — Completion, Chat, Code Review et Agentic — to deliver smarter, more context-aware assistance through continual model refinement et in-context learning.
Key research et engineering challenges include:
-
Elevating LLM quality for developer-centric tasks via iterative training, feedback loops et real-world usage signals.
-
Harnessing code structure et context— leveraging syntax, semantics et repository context enrich both training et inference.
-
Scaling responsive code generation to enable fast, interactive experiences for millions of developers.
I thrive at the intersection of cutting-edge research et large-scale deployment, turning AI breakthroughs into tools that make every developer more effective.