About the role
We need someone who’s built agents with their own hands — not managed AI products from a roadmap. If you’ve shipped an agent that actually works, or spent weekends wiring LLMs into tools that solve real problems.The Role Own the agentic AI product line at a supply chain AI SaaS startup. End-to-end. Product definition, build, ship, iterate. Small team, high autonomy, direct line to revenue.
MUST-HAVES
- You’ve built at least one working AI agent or AIpowered tool — personal project, hackathon, startup, side hustle, anything. Show us.
- You use AI tools daily in your own work — Claude, Cursor, GPT, Copilot, Claude Code, Replit Agent, or similar. Not as toys. As productivity multipliers.
- You can write prompts that actually work. You understand system prompts, chain-of-thought, tool use, structured outputs.
- You think in systems — how agents compose, where trust boundaries go, what happens when an agent fails, how humans stay in the loop.
- You ship fast and iterate. V1 mindset. You’ve built something in a weekend that people actually used.
- You can talk to enterprise buyers without sounding like a tech bro. Empathy for the planner who’s been doing this in Excel for 20 years.
STRONG SIGNALS (NOT REQUIREMENTS)
- You’ve worked with the Anthropic SDK, LangChain, LangGraph, CrewAI, AutoGen, or similar agent frameworks.
- You’ve built with Python + Streamlit, or shipped a React/Next.js prototype that a real user touched.
- You understand supply chain concepts — or you’re the kind of person who’ll learn demand planning in a weekend and show up Monday with a POV.
- You’ve done product work at an early-stage B2B SaaS company.
- You’ve built MCP servers, Claude Code skills, or contributed to open-source AI tooling.