JOB-1619
A seed-funded startup that builds AI agents to walk prospects through product demos is growing quickly and wants founding engineers to shape its technical culture. Engineering is currently tiny, so the people hired now will set the standard for everyone who joins later.
What the job looks like
- Debug live customer agents by reading conversation logs, finding where the language model went wrong and shipping orchestration fixes the same day
- Prototype adaptive, generative interfaces that rearrange themselves around what each prospect cares about
- Maintain the quality-measurement pipeline, trace dips to their cause, revert bad changes and add tests
- Design infrastructure such as agent versioning, letting clients try out modifications ahead of release
- Cover the customer-facing control panel plus voice, low-latency serving and instant starts
The right person
- Has built systems that people actually depend on and can describe how they broke in practice
- Brings sound judgment about when to ship and the discipline to roll back or retire features that fail
- Understands the distance between an LLM prototype and a dependable product, and closes it with evals and iteration
- Is comfortable across the full stack, using React and Python, with strong front-end and product instincts
- Takes initiative without waiting, and accepts small tasks like CSS fixes and migration scripts alongside the interesting problems
- Has a few years at a venture-backed B2B software startup, ideally with customer-facing AI agents shipped to production
On-site in the San Francisco Bay Area. No visa sponsorship or transfers; candidates need existing U.S. work authorization.
