Joseph E. Gonzalez
Confidence 0.75 · 2 sources · last confirmed 2026-09-04
Professor of EECS at UC Berkeley and co-director of the Sky Computing Lab; a principal investigator behind the Gorilla project (LLMs connected to APIs) from which the Berkeley Function Calling Leaderboard grew, and a co-author on the lab’s line of work on serving and evaluating LLM systems.
Promoted on the second-source rule, alongside Ion Stoica, as senior author on both of the corpus’s Berkeley agentic-evaluation papers.
Appears in this wiki via
- 2025-07-13-patil-berkeley-function-calling-leaderboard — senior author on the BFCL paper (ICML 2025); the Gorilla project’s evaluation instrument.
- 2025-03-17-cemri-why-do-multi-agent-llm-systems-fail — senior author on MAST, the multi-agent failure taxonomy (NeurIPS 2025 D&B).
Both papers converge on the same diagnosis from opposite methods — see multi-agent-failure-modes and ai-benchmarks.