Artificial Intelligence Engineer

DocsDocs is built AI-first, in what we ship and in how we build it. Coding agents and automated review pipelines are how a small team keeps this codebase moving fast.

We're hiring AI Engineers to push that further: design and maintain the AI development workflows the whole team builds on, from coding agents and custom tooling to automated testing and documentation pipelines. We handle protected health information, so doing this safely matters as much as doing it fast. You'll think hard about what AI tools can touch, what they can't, and how to keep agent workflows secure in a regulated codebase. The role includes improving the team's AI tools and shipping features with them.

What you'll do

• Own the team's AI workflows. Design, build and maintain the coding agents, automated review pipelines, doc generation and test scaffolding the whole engineering team relies on every day.
• Build development tools. Write custom dev tools and MCP servers that connect our agents to the codebase, CI and tickets, safely and reliably.
• Set the guardrails. Decide what agents can and can't touch in a HIPAA-regulated codebase, and keep PHI and secrets out of model context by design.
• Ship product as well as infra. Use the workflows you build to ship product features and test them in daily development.
• Own model quality. Evaluate and integrate model APIs (Anthropic, OpenAI, Google); build evals, track regressions, and keep prompt output reliable as things change.

What we are looking for

Real experience building AI workflows: coding agents, automation pipelines, or multi-step LLM tooling that you or others actually used.

• Strong prompt engineering instincts: you know how to get reliable, repeatable output from a model and how to debug it when it drifts.
• A solid foundation in hands-on software development beyond AI tooling, ideally on the web side (TypeScript, React, Node, or similar); the workflows you build still have to be good code.
• Good judgment about when AI helps and when it gets in the way.
• Daily fluency with AI development tools (Claude Code, Cursor, Copilot, or similar). These are core to how we work.
• Comfortable working with at least one major model API (Anthropic, OpenAI, or Google).
• A track record of using AI to ship faster: side projects, work projects, or open source all count.
• Awareness of the security side of AI tooling: agent permissions, prompt injection, and keeping secrets and sensitive data out of model context.