Lab Overview
Transmute Data is an AI research and engineering lab that works inside your team.
Our consultants take on client problems as forward-deployed senior consultants. Separately, the lab builds its own consulting agents, models, and forward-deployed consulting frameworks.
What we publish about them is in Publications.
An industrial research lab
Research and engineering under one roof
Transmute Data works in the spirit of the great industrial research labs. Research and engineering sit under one roof: the people who hold the research record also write the code, and senior consultants lead the work.
It is one lab whose researchers embed with different clients. It builds its own models, agents, and tooling, and it publishes what it can openly, in Publications.
How the lab works
Senior consultants lead, start to finish
Senior, industry-leading consultants run every engagement and stay on it: twelve data scientists, AI engineers, and researchers, averaging 15+ years each. The people you meet are the people who lead the work, from the first call to handover.
What we own
Our own tools, never trained on your data
We build our own designs, agents, harnesses, skills, and processes, and improve them using our own internal development work. Never customer data, never PHI. We hold no HIPAA attestation of our own — we work inside your governed environment, under your BAA.
How we handle your data is set out on one page: Security and data handling.
Ownership
Yours once paid; our tools stay ours
You own the deliverables once they're paid for. We keep the tools and frameworks we brought with us. Those include our consulting agents, models, and forward-deployed consulting frameworks.
Publications
What we publish
Our consultants publish under the Transmute Data brand on the Publications page.
Infrastructure and frameworks
How our tools work
Notes on the infrastructure and frameworks Transmute Data owns, and how they work inside.
Agents and models
What the lab builds
Notes on our agents and our own language models.
Open source
Code we release
Our open-source contributions, with links to the code.
Client work, anonymized
Deliverables, with names removed
Client deliverables with client names and data removed.
Everything here is published by our consultants under the Transmute Data brand.
Latest from Publications
Agents and models
Agentic Operation, in plain language
Agents and models
Agent anatomy for a healthcare workload ladder
Agents and models
Device-ladder eval figures
A 30-minute call, on one problem you name.
Shaun takes the call. You leave knowing whether someone on our team has solved that shape of problem before, and who.