Open role
LLMOps Engineer
You own the difference between a model that demos well and a system that holds up under load, under budget, and under audit. Evals, observability, cost, latency, and the failure modes nobody writes down.
What you'll do
- Build the eval harnesses that tell us whether a change helped
- Instrument production agent systems — cost, latency, failure modes
- Tune the cost/quality trade-off deliberately rather than by accident
- Support on-premise and data-residency deployments where a client requires them
What we look for
- LLM systems in production
- evals and observability
- cost/latency tuning
Nice to have
- RAG at scale
- on-prem / data-residency deployments
- guardrails