Services

What we work on

We take on a small number of problems at a time and see them through, from the first design to something reliable in production.

Applied AI

Retrieval, embeddings, and LLM features built into real tools. We focus on systems that are useful day to day and honest about what the model can and can't do.

Distributed systems

Services that stay correct under concurrency and load. We design for failure, keep state consistent, and make the behaviour easy to reason about.

Developer tools

CLIs, libraries, and workflows that respect your time: fast, scriptable, and happy to run on your own machine, in the spirit of the Unix philosophy.

Tech stack

What we build with

We are not loyal to any one tool. We reach for what fits the problem and what we can keep reliable in production. A fairly complete map of what we work with:

RAG
Agents
Pipelines
MCP
LangChain
LangGraph
Protocol Buffers
OpenTelemetry
Vector search
Event-driven
System design
Agile delivery

Start a conversation