Founding Engineer

Founding Engineer NYC (Manhattan, on-site) Pre-seed - $3M raised$220k + 1.5% equity Build the interface agents actually use. This is a pre-seed startup (funded by well-known early-stage VCs) building the CLI and API layer that lets AI agents pull B2B data reliably — 80+ providers, one install, no dashboards, no seat licences. Think: the plumbing between Claude Code / Cursor and the data those agents need to actually get work done. You'd own: • The CLI & SDK — install flow, auth, command structure (Python core, TypeScript web layer) • AI agent tooling — MCP integrations, skill docs, tool interfaces • The data engine — rate limits, credential rotation, cost optimisation across 80+ providers • The provider network — every integration has its own quirks, you'll build and maintain the system that handles them First 90 days: ship your first provider integration in month one, own a full chunk (credential management or an MCP integration) by month two, and by month three you're building the core agentic optimisation framework — the infrastructure that lets agents reason about which skills to use and when to backtrack. Your Profile: • 3+ years in fast-paced startups — founding engineer or early-team experience • You've built dev tools before (CLIs, SDKs, APIs) and know what good DX feels like • You've fought third-party API pain (rate limits, flaky auth, inconsistent schemas) and won • Recent hands-on exposure to agentic systems / AI tooling • Clear, ego-free communicator — you'll be talking directly to founders and real users • Comfortable with ambiguity — this is zero-to-one, not a roadmap handed to you Comp: $180K–$220K base + 0.5–1.5% equity (founding-team grant), health/dental, on-site in Manhattan. Process: no LeetCode. Intro call → two system design screens → a paid half-day working session with the team. Not looking for the pedigree-school-to-FAANG path — looking for someone who's taken something from zero to one and still has the scars to prove it.