Spring 2027 · New England · Tentative
Agent Evaluation Science
Summit
Two days on how agentic AI systems are designed and evaluated, for everyone who builds, deploys or studies them.
- Date
- Mid-April 2027
- Location
- Kendall Square · Cambridge, MA
- Scale
- Day 1 up to 500 · Day 2 150–200
Overview
A research day, then a working day
Everyone who builds or deploys agentic systems has to design and evaluate them. The summit brings groups from across New England together to compare how they do it, surface the shared problems and disagreements, and leave with questions that lead to follow-up research, replications or shared infrastructure. Day 1 is open to up to 500 people; Day 2 is a working day for 150–200 people, in working groups and open discussion.
Program
A research day with keynotes, a panel, talks and posters, then a working day of working groups and open discussion.
View program →02Take part
Present your group's evaluation work or submit a talk or poster. The call opens in January.
How to take part →03Organizing team
Five working responsibilities, each with an owner who recruits their own helpers.
Open responsibilities →04Sponsors
Sponsorship pays for catering, recording, posters and travel support, and keeps tickets affordable.
Sponsor the summit →