AI Agent Developer

What we do We build AI agents for business professionals — real deliverables in production for paying clients, not demos. Clients are happy and demand is growing faster than one person can build. Long-term, we're developing a standardized framework for how our agents get built, so this role can grow into shaping that. What you'd actually do Take over and extend production agent projects (current stack: Node.js, Vercel, agentic workflows, third-party AI APIs including video generation — no prior experience with any specific vendor required, you'll pick them up) Debug and un-stick hard problems: flaky vendor APIs, workflow orchestration, behavior that "works but feels wrong" Own technical decisions — and explain them, in plain language, to a founder who is technical enough to follow you but is not a developer. If you can't explain a tradeoff without jargon, this isn't the right fit. Work with AI tools openly. We build with Claude and expect you to. What we care about is your judgment about what the AI got wrong. You're probably a fit if You've shipped and maintained production systems — things real users depended on, where you handled what broke after launch You've built LLM-based features or agents beyond tutorials: tool calling, workflow orchestration, deciding what belongs in the model vs. deterministic code You're a strong generalist. We'd rather have someone who learns a new vendor API in two days than someone who memorized one stack You communicate proactively in writing — short updates, honest about unknowns, surfaces problems early You're probably not a fit if You need a full spec before you can start Your first instinct on inheriting a codebase is "rewrite it" You treat AI output as done without verifying it — or refuse to use AI at all