Senior Platform Engineer

About the role

The role

Own the systems behind Coval’s evaluations. Coval is the evaluation platform for conversational AI: we help enterprise teams simulate before launch, observe in production and review with humans. This role owns the machinery that makes all of it run.

That means the distributed pipelines, queues and orchestration that score tens of thousands of conversations a day, plus the data layer and cost controls underneath them. You’ll own the systems enterprises run their voice agents through, and you’ll get in while they’re still being built.

What you’ll do

- Build out our distributed execution engine: the workers, queues and orchestration that fan out tens of thousands of simulated conversations a day and bring back scores

- Own the data side, from the databases holding conversations, audio and judge results to the pipelines that move them between stages

- Make real cost trade-offs. Every model call and storage byte has a price, and you keep unit economics sane as volume climbs

- Stand up and maintain the environments where our LoRA training jobs run, as a key part of the plumbing for our ML work

What we’re looking for

- 4+ years building real systems, with Python and AWS as home turf

- Experience with managed services and Terraform, which we use heavily, and the judgment to know when a managed service stops being the right answer

- You’ve operated something under load and know what it feels like when a queue backs up at 2am

- You can look at a half-built system and decide what to finish, what to fix, what to start next and what to buy time on

- High agency: on a small Series A team, nobody writes your tickets. You figure out what matters next and take it there

- Ambitious and kind. As a founding member of the engineering team, you’ll shape the architecture and the standards everyone else builds against

See the posting on Coval's site