Learns on the job
Every task, failure and correction becomes a lesson. Your AI improves from its own traffic, with little labeled data needed.
With Noeri, your company builds specialized AIs from its own data and know-how.
They keep learning from real work, team up on your hardest problems, and remain entirely yours.
Our vision
Read our visionImagine intelligence that grows the way life and culture always have: in countless forms, built and owned by those who know their problems best. The alternative is a single superintelligence, owned by a few, whose blind spots become the world’s blind spots.

Our differences are the source of our collective intelligence, and much of what makes the world beautiful.
Our product
Noeri works alongside your teams on the agents you run today and the ones you set out to build. Those agents keep improving from real work and grow with you toward the hardest problems in your field.
Every task, failure and correction becomes a lesson. Your AI improves from its own traffic, with little labeled data needed.
Not just prompts. Your AI refines its skills, tools and orchestration as it works, and creates new ones when the existing ones fall short.
Your AI doesn't just pick models, it trains its own on your work, so each part of the job runs on the model that does it best.
Quality and cost are measured on your own traffic and metrics, not a public benchmark, so the improvement is visible.
Our research
Our team created some of the most widely used tools and methods in machine learning. Now we’re building AI that improves itself at every layer, from data and tools to the models themselves, with every experiment run in production.
Explore our researchCreate
Can the craft of building AI be learned? We build systems that can optimize themselves, generate their own data and curricula, and create their own tools.
Own
How little data and compute does it take to master a domain? We work on distillation, lightweight updates and learning from limited data.
Evolve
How can AI keep learning after deployment without forgetting? We work on learning at test time and from rich feedback.
Discover
How does AI find answers no one has found? We work on search, reasoning, and teams of models that propose, test and refine.
The future of many intelligences is built by many creative minds.
Our founding team
, , , , , , , , , , , , , , , ,
Prithviraj AmmanabroluProf at UCSD. Language agents via RL for over a decade across academia and multiple startup exits. Enjoys boba tea.

Co-CEO
Yejin’s career has been anything but linear, a prime example of continual learning on the job — the very essence of Noeri.
Along that winding road, from commonsense reasoning to pluralistic alignment to her more recent passion for AI for science, she has been recognized as a MacArthur “Genius” Fellow, a two-time TIME100 AI honoree, a TED main-stage speaker, and the recipient of more than a dozen best-paper and test-of-time awards. Today she is at Stanford as Professor of Computer Science and Senior Fellow at HAI. She previously led research at Ai2 and NVIDIA, where she acquired a taste for risk-taking and shipping research into products.

Co-CEO
If you’ve built a machine learning model, there’s a good chance you’ve used Carlos’s work.
He co-created XGBoost, one of the most widely used tools in machine learning. Noeri is his seventh startup; he loves to invent, build and ship. He co-founded Turi, which Apple acquired in 2016, then led machine learning and AI at Apple. Today he is a Stanford professor, Director of the Stanford AI Lab (SAIL) and a member of the National Academy of Engineering, with 16 best-paper and test-of-time awards. His research on continual learning, AI that keeps improving after it ships, is the heart of Noeri.
Join us
Our team brings together researchers and engineers from leading academic institutions and technology companies. We work as one collaborative team, where research and engineering move together, and every idea is tested in the real world.