Welcome to Exa. Exa is an applied AI lab building a perfect search engine unlike the world has ever seen.
We spent years building massive-scale indexes and researching new retrieval systems for the agentic world we knew would arrive. That world has now arrived.
Today, Exa is the highest quality search API at every latency. We power search for Cursor, Cognition, HubSpot, OpenRouter, Monday.com and over 400,000 developers.
In the next few years the number of searches from AI agents will be 1000x more than Google searches today. Before us lies an opportunity to reinvent how the world gets information.
Information is critical civilizational infrastructure for our new AI reality. Politics is fragmenting, wars are raging, and AI is accelerating — we need tools we can trust that deeply inform us what's going on. If we can achieve perfect search so that every AI has the highest quality information, then every human will too.
Will Bryk
Research
We're an applied AI research lab organizing the entire web
Observability: Since the runtime sits at every invocation boundary, it records timing, inputs, outputs, and decisions automatically and sees everything by default. It enriches any error a node throws with the context of which node failed and what was happening upstream.
In multi-turn agentic RL, higher per-step success makes rewards less sparse and helps the model learn from fewer rollouts. This may be due to the semantic nature of Exa search, which is designed specifically to handle non-keyword natural-language queries.
For role based discovery, an LLM judge is used given there is no single correct answer. For each result, we fetch the page content and verify whether it matches the query criteria. A search for "senior software engineer in SF" for example returns 9 matching profiles and 1 person in San Diego scores 90%.
Search is relevant precisely where this parametric memory fails: for tasks that are new, niche, or complex. For agents writing code, relevant documents such as changelogs, GitHub issues, and SDK docs are continuously updating, making real-time fluency a necessity.
We find that training with Exa requires fewer tokens for the same number of steps.
We reformulate it as a discriminative task. Instead of "generate the right answer from this context," we ask "does this context contain the right answer?" A frontier model can reliably discriminate whether a highlight contains a given fact, even when synthesis would be noisy.
Will was one of the first engineers at Cresta building real-time AI products, after studying CS and physics at Harvard.
Co-founderJeff Wang
Jeff spent 3 years building data and web infra at Plaid after studying CS & Philosophy at Harvard, where he ran a GPU cluster in his dorm room.
Technical StaffBen Chen
Ben previously did quant trading at SIG after taking the hardest math course in the country at Harvard.
Technical StaffHubert Yuan
Hubert previously worked on projects like particle simulations and automated wheelchairs, after studying CS in the Yao Class at Tsinghua University.
Technical StaffShreyas Sreenivas
Shreyas built projects ranging from training neural networks in Haskell to a game streaming engine, after studying CS at Waterloo.
Technical StaffMichael Fine
Michael previously worked on ML and privacy at various companies, including Apple, after studying CS at Harvard University.
Technical StaffFelix Zeller
Felix built open source tools spanning next-gen text editors and knowledge management systems, after studying CS and philosophy at UIUC.
People + Workplace OpsStacey Tara
Stacey previously worked on business development at Awesomic, after earning her bachelors and masters in Kyiv.
Technical StaffJoshua Ahn
Joshua studied CS at UChicago working on ML for 3D reconstruction, and now builds virtual worlds and massive datasets at Exa.
Technical StaffIshan Goswami
Ishan previously cofounded a text-to-video startup, and before that was a software engineer at Rephrase AI, which got acquired by Adobe.
Technical StaffJoao Adriano
Adriano skipped college and has been coding professionally for 12 years, most recently hacking equipment protocols on cargo ships.
Technical StaffCarlos Marques
Carlos did MLOps for years, building training infra for thousands of models across cloud and edge, after studying engineering in Portugal.
Technical StaffGabriel Cammany
Gabriel founded Sofon, a graph-based people-ranking tool, and was a founding engineer at Diagonal, an API for crypto payments.
Technical StaffTom An
Tom studied CS at CMU implementing network protocols and compilers, after working at Google on highly optimized, global-scale services.
Chief of StaffFelicia Tang
Felicia previously worked on Growth/Strategy at startups in search and climate tech, and prior to that did private equity at Blackstone.
Technical StaffAnca Negoiu
Anca previously built AI search pipelines and databases at Primer.ai, after studying Computer Science at Princeton.
Head of R&D Product EngineeringLiam Hinzman
Liam previously worked as a software engineer at Numerai before becoming an artist and joining Exa to lead our R&D product team.
Technical StaffSam Mitchell
Sam studied CS at MIT and then researched efficient GNN architectures at the IBM Watson AI Lab.
Design EngineerSong You
Song spent two years before Exa learning how to illustrate and paint concept art, after studying Computer Engineering at University of Toronto.
ProductShyam Kumar
Shyam founded Kapstan, a Kubernetes abstraction layer for container orchestration, after engineering at Apple and consulting at BCG.
Technical StaffMark Pekala
Mark worked on video at Ameelio and Kino AI, and data materialization at Sigma Computing, after studying math and CS at Harvard.
Technical StaffTyler Killian
Tyler studied math and CS at McGill University, where he scored Top 500 on the Putnam and was a Codeforces Candidate Master.
Technical StaffLinh Nguyen
Linh previously worked in fintech on explainable ML models for credit underwriting, after studying Chemical Engineering at MIT.
SalesWill Roberts
Will studied Industrial Engineering at Michigan and Berkeley, previously a research engineer at SRI International.
Technical StaffBen Chan
Ben did a PhD at Cornell building some of the fastest distributed algorithms in the world, now taught at universities worldwide.
Technical StaffRohit Prakash
Rohit graduated in CS from NYU debugging kernel programs, after a coding journey that began at age 11 making games in C++.
Technical StaffMax Buckley
Max spent 12.5 years at Google across Europe on finance, trust and safety, ML, information retrieval, and large language models.
Investors
Backed by world-class investors
In May 2026, Exa reached a $2.2B valuation, backed by some of the world's leading venture firms and angel investors. We're grateful to have these incredible partners alongside us.