Build primitives that last.
Small team. Hard problems. Open source. We work on systems-level infrastructure for AI agents — Rust, TypeScript, information retrieval, knowledge graphs. We ship in public, write specs before code, and optimize for correctness over speed-to-market.
If you want to build the primitives the next generation of agents will depend on, and you care about getting the details right, we should talk.
The culture.
Specs before code
We write design documents and get alignment before opening an editor. The spec is the product — code is the implementation.
Ship in public
Our core is Apache 2.0. PRs are reviewed in the open. We write changelogs that explain why, not just what. If it is not in the repo, it did not happen.
Correctness over speed
We would rather ship next week with tests than today without them. Type systems, property-based testing, and formal reasoning are tools, not overhead.
Deep work, async-first
No standing meetings. No Slack pings for things that can be a GitHub comment. Long blocks of uninterrupted time are the default, not the exception.
Introduce yourself.
We are an early team based in Bangalore. We do not have specific reqs posted right now, but we are always looking for exceptional people.
We are especially keen to talk to systems and infrastructure engineers who are strong in Rust and distributed systems — people who want to own the hot path from memory ingestion to retrieval, the storage engine, and the deployment story. If that sounds like you, send your GitHub, a project you are proud of, and a few sentences about why Arc Labs to careers@arc-labs.ai.