Applied AI
AI Engineer
Build agents that can understand, decide, and act reliably inside real hospitality operations.
The mandate
Move AI from impressive demos to dependable operational systems with measurable learning loops.
The exact shape will evolve. We care more about the outcome you can own than preserving the boundaries of a job description.
The work
What day to day actually looks like.
The mix changes with the company’s bottleneck, but these are the recurring loops you should expect to own.
- 01
Read traces and real customer failures to identify the next reliability bottleneck.
- 02
Build and test agent tools, orchestration, memory, prompts, and evaluation harnesses.
- 03
Work with product engineers to place AI inside a clear operator workflow, not beside it as a demo.
- 04
Measure quality, latency, and cost in production and turn incidents into durable eval coverage.
The outcomes
What success starts to look like.
- Deploy an agent capability that reliably completes a real operational job.
- Build an eval and trace loop that catches regressions before customers do.
- Improve a production quality metric while keeping latency and cost within useful bounds.
The person
Signals we look for.
You do not need to match every line. Show us evidence of slope, ownership, and the quality of judgment you bring to unfamiliar problems.
Could you own this at Trellis?
Apply with evidence of what you have built, changed, or learned unusually fast.