What the Etheon Lab works on
The Etheon Lab studies what makes AI software safe to ship: evaluation, guardrails, observability, safe releases and governance.
- Publication

On this page
The Etheon Lab is where we research the engineering behind the AI software we build for companies. This note sets out what it works on.
1. What the Lab studies
- Evaluation: how a model is tested before it reaches users, from the quality of its answers to its safety and its drift over time.
- Guardrails: limits that hold whatever a model answers, such as safety filters, tool permissions and a person approving sensitive steps.
- Observability: telemetry, lineage and drift detection that keep a model accountable once it is live.
- Safe releases: shadow and canary deployments, staged rollouts and rollback, so any change can be undone.
- Governance: audit trails, model cards and versioned releases that show what changed and why.
2. What it publishes
Each piece is filed by kind: a publication, a milestone, a conclusion or a release. Where a study shapes how we build, it shows on our service pages, as the release gates do on the AI page.
3. Products and open source
The Lab also builds our own product, Etheon Delta, and publishes open-source code on GitHub for engineers to read and run.
Related research. More from our Lab on AI.
How we build an AI agent, in five steps
An agent is software that takes actions in your systems. We build it in five steps, inside limits held outside the model, with a person where it matters.
Which AI model should we use?
The one that does your task best at a cost you can carry. We score candidate models on your evaluation set and choose by quality, cost and response time.
How we test a model before release
Every model we build passes the same release gates before it reaches users: task quality, grounding, safety, privacy, latency, cost and drift.