Joan Cabezas works on AI agents, post-training and evaluation. His projects include Terminal-Bench 3, SWE-Marathon, computer-use infrastructure and knowledge-graph research.

Claimed by the subject

Joan Cabezas works on AI agents, post-training and evaluation. His projects include Terminal-Bench 3, SWE-Marathon, computer-use infrastructure and knowledge-graph research.

joan.so

Important Problems

Joan Cabezas maps open questions in AI research, including generalization, memory, evaluation and sustained work on long-term objectives.

Open ↗

Joan's Collections

Joan's Lists

In these lists

Cosigns Given