I build AI agents and multi-agent systems.

Currently, I work at Alibaba’s Tonyi Lab, based in Bellevue, USA as a Senior Staff Engineer. I lead the development of production-grade agent systems and moonshot projects to advance the state of the art in AI agents.

Previously, I was a Principal Researcher at Microsoft Research based in Redmond, USA. I was the lead architect of AutoGen , an open-source framework for building AI agents and multi-agent applications. AutoGen provides a simple, high-level API for creating and managing agents that work together to solve complex tasks autonomously or under human supervision. It also provides an event-driven, low-level API for building workflows and gives developers full control of agent behavior. The framework is based on the Actor Model of distributed computing, with an agent runtime layer that can be deployed in a distributed environment, hosting agents created with different programming languages. AutoGen has been widely adopted by developers and researchers inside and outside Microsoft. I was fortunate to lead a vibrant community of 500+ contributors, many of whom added major new features and capabilities to the framework.

During my time at Microsoft, I also joined forces with Azure AI to build the Foundry Agent Platform, a cloud platform for building and deploying AI agents and workflows at scale. As part of this effort, I helped build the Microsoft Agent Framework , as an evolution of AutoGen for enterprise-grade applications and a core component of the Foundry Agent Platform.

Prior to working on AI agents, I worked on database systems at Microsoft Research. We developed a cost-based, platform-independent query plan rewrite rule for MATCH_RECOGNIZE queries in general-purpose SQL engines. The new rule boosts median query latency by 5.4X in Trino. In an even more ambitious undertaking, we developed a specialized execution engine for MATCH_RECOGNIZE, which includes an extended set of operators and a query optimizer based on a novel cost model. Our work resulted in a 6X median performance improvement over state-of-the-art specialized execution engines.

Before joining Microsoft Research, I completed my PhD in Computer Science at the University of Toronto, under the guidance of Prof. Renée J. Miller. My thesis is in dataset search over massive Open Data archives. Specifically, I contributed algorithms for large-scale set similarity search and data sketches . These algorithms can find joinable or unionable tables from over 100K tables in milliseconds. Based on my research work, I built an Open Data search engine stack to make it easy for people to use Open Data in their applications.

I am constantly exploring new ideas in computing through my open-source projects and writings. My goal is to work towards the democratization of computation, whereby advanced A.I. and algorithms are easily accessible to everyone.

[Github] [Google Scholar] [Blog] [Email]