Bio
Lakshya A. Agrawal is the creator and maintainer of GEPA and a second-year EECS PhD student at UC Berkeley’s Sky Computing Lab. His research focuses on optimization, evaluation, and self-improvement for LLM-based agents and systems. He previously worked as an AI4Code Research Fellow at Microsoft Research.
Session (1)
Additional Speaker Details
Using publicly available information we constructed an analysis to help you get a feel for this speaker before deciding to attend their session.
Lakshya Agrawal is the lead author of GEPA (Reflective Prompt Evolution Can Outperform Reinforcement Learning), an ICLR 2026 oral showing a prompt optimizer that beats GRPO by ~6pp on average (up to ~19-20pp) while using up to 35x fewer rollouts, and the creator of the open-source GEPA framework (5.4k stars, 50+ production deployments cited). An AI PhD student at UC Berkeley and former Microsoft Research Fellow, he is also first author of optimize_anything, a universal API for optimizing any text parameter. His session is a chance to hear directly from the researcher who built and shipped one of the most-starred open-source prompt-optimization frameworks.
Speaking style
Recent talks (3)
GitHub
@LakshyAAAgrawalRecent writing (2)
- GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning · paper · Jul 2025
- optimize_anything: A Universal API for Optimizing any Text Parameter · paper · May 2026
Podcasts & interviews (2)
- GEPA with Lakshya A. Agrawal - Weaviate Podcast #127! · Weaviate Podcast · Aug 2025
- Ep. 52: Lakshya Agrawal, GEPA Lead Author and Berkeley AI PhD Student · Delta Podcast · Nov 2025