LangWatch is the AI agent engineering platform, and the open-source frameworks behind it, that developers use to ship reliable agents with confidence.
We saw the same problem from two sides: agents that demo beautifully and then break in the messy reality of production, with no systematic way to catch it. Testing AI was slow, manual, and owned by whoever happened to be closest to the incident.
LangWatch is the platform we wished we had: simulate real conversations before launch, evaluate every release, and watch quality in production, so shipping reliable agents is the default, not a heroic effort.
Make AI reliable enough that teams can focus on strategy and creativity, while we help ensure their AI behaves as expected.
Commercial leadership at Lightspeed and Biller before LangWatch. Met Rogerio at Antler Amsterdam, where the GenAI side-projects they were both working on turned out to share one missing piece: quality control. Runs go-to-market and the team.
Spent years at Booking.com, including helping build its first AI agents and seeing firsthand how fragile, opaque, and hard to test these systems become at scale. Engineering background at ThoughtWorks before that; full-stack across React, TypeScript, GraphQL, and ML.
Plus a hands-on team of engineers, researchers, and operators across EU and the Americas, hiring for the people who want to build the reliability layer for AI from scratch.
We picked these because we needed them to make decisions every day, not because they look good on a page.
Every decision starts with the team shipping AI and the people they serve.
Bias to action. We would rather ship, learn, and correct than wait for certainty.
Engineers, product, and domain experts own quality as one team, not a hand-off.
Automate the repeatable so people spend their time on the hard, human parts.
Tomorrow brings new challenges. We close the loop on the ones in front of us now.
Start free in minutes, or talk to us about your agents and where they need to be more reliable.
