Opening keynote · Day 2
About Amara
Amara leads a 60-person research group building evaluation methods for large language models. Her team's open benchmark for reasoning is used by most major AI labs.
Before Northwind she spent eight years at a university robotics lab and co-founded a startup that taught warehouse robots to learn from demonstration. She is known for talks that make hard ideas feel obvious.
The most important question in AI isn't 'can it do the task?' but 'how would we know if it couldn't?'
40+Papers published
12Years in AI
3rdTime at Vector
Sessions at Vector
- Opening keynote: Measuring what matters in AI
- Panel: Open models, open questions
- Office hours: Evals clinic
Background
- PhD Machine Learning, Imperial College
- Co-founder, Handover Robotics
- Advisor, Open Evals Alliance
Send a question for the Q&A
More speakers
