Research on AI systems that do the work.
How we build, test, and improve Lily across enterprise systems, agents, and long-running work.
Company · July 2026 · Series B
The Engineer That Never Leaves
Announcing our $60 million Series B led by Energize Capital, and the belief behind it: the model is no longer the bottleneck in enterprise AI. Deployment is. Lily learns how a business works, builds the software it needs, deploys it inside the systems already running the business, and stays.
Read the announcementEngineering Notes · No. 01-02 · July 2026 · Two-part series
Build the Environment, Then Optimize the Harness
How we turn real production codebases into a private, resettable evaluation environment, and then point an automated optimizer at it. Part 1 builds the replica and reports what it says about Lily, our coding agent. Part 2 freezes the model, evolves the harness around it, and shows how to tell a real improvement from a lucky one.
Read the series