Weiran Yao

I'm a research scientist at Salesforce AI Research in Palo Alto, where I mostly work on development of multi-agent AI systems, and LLM post-training for Agentic AI.

At Salesforce I've led the development of Agentic AI systems including DigitalHQ, SEDA, Salesforce Research Ensemble Agents, and Salesforce WebAgent.

For LLM training, I've led post-training of Retroformer, a general critic model for agent self-reflection. I've worked on xLAM, a family of state-of-the-art language models for function calling. I conducted research on synthetic data generation pipeline for APIGen and AgentOhana.

For Engineering Products, I developed automatic root cause analysis algorithms, or AI for IT Operations. I worked on function call and structured output API endpoints for Salesforce xLAM services. I did my PhD and MS in Machine Learning at CMU, where I was advised by Kun Zhang.

Email  /  CV  /  LinkedIn  /  Scholar  /  Twitter  /  Github

profile photo