RAG in the age of agents. SWE-Bench as a case study.

Jason has built search and recommendation systems for the past 6 years. He has consulted and advised a dozens startups in the last year to improve their RAG systems. He is the creator of the Instructor Python library.

I was previously a founding member of technical staff at Augment Code, a leading AI coding company building everything from completion models to agents, all with deep awareness of your company's codebase and knowledge graph. We raised at a ~$1B valuation in early 2024. Prior to this, I was a researcher at Facebook AI Research, where I studied AI reasoning in the context of board games. I co-authored a paper in Science where we built an AI that reached human-level in the strategy game Diplomacy.