Product Engineer

Job Description

We build AI-first. Our engineers direct AI agents through most of the code they ship — planning with them, reviewing their output, and owning the result. If Claude Code, Cursor, or a similar agentic workflow isn't already central to how you work, this role won't be a fit. If it is, you'll feel at home on day one. We're building agentic AI products that own entire workflows, not just answer questions. Our two products: Ozyn , an AI assistant that executes tasks across Gmail, CRM, Calendar, Sheets, and 50+ integrations — and Alvoff , a conversational procurement platform with agentic RAG and supplier intelligence serving buyers across 20+ industries. We practice what we build: our own development is AI-first, end to end. What you'll do You'll own features end-to-end across Ozyn and Alvoff: talk to users, scope the problem, build it, ship it, and watch how it lands. There are no handoffs between \"product\" and \"engineering\" here — you are both. Ship full-stack features across [your stack, e.g., TypeScript/React/Postgres] with AI agents doing the bulk of implementation under your direction Talk directly to users and turn ambiguous problems into shipped product, often within days Set the bar for what AI-first development looks like — better prompts, better agent workflows, better review practices — and raise it as the tools evolve Make pragmatic calls on quality: knowing when agent output is good enough to ship and when it needs your hands on the keyboard Contribute to product direction, not just implementation. If you think we're building the wrong thing, we want to hear it How we work Agents write most of our code. Your job is judgment: decomposing problems, directing agents, reviewing critically, and owning correctness. Typing speed is not the bottleneck; taste and decision speed are. Small team, high ownership. You'll ship to production in your first week. Users over process. We optimize for learning fast, not for ceremony. What we're looking for Strong software engineering fundamentals — you can evaluate and correct AI-generated code because you could have written it yourself A genuinely AI-native workflow today, not aspirationally. You have opinions about agent orchestration, context management, and where these tools break Product sense: you've shipped things users loved (or hated — and you learned why) Comfort with ambiguity and speed. You'd rather ship a v1 this week than a v3 next quarter 3+ years building production software, or a portfolio that makes years irrelevant Nice to have Experience at an early-stage startup or as a founder You've built with LLM APIs (evals, agents, RAG) — useful, but note this role is about building with AI, and only sometimes building AI features Public evidence of how you work: repos, posts, tools you've built for yourself