Founder's Engineer

Founder's Engineer

San Francisco, onsite hybrid only Full-time

San Francisco, hybrid (onsite hybrid only) · Seed-stage · Building the Agent OS for business

What this actually is

Most founder's-associate roles turn into coordination and status updates. This one is different: you build the systems that make status updates unnecessary, and you build them on the product itself.

We are building an operating system for business agents: permissioned agents running on dedicated infrastructure, with every execution signed and auditable. You will run the company on it. Recruiting, GTM, operations, reporting: wherever a human is doing repeatable work, you design the agent that does it, ship it on our own platform, and own the result. The talent engine is a concrete example: sourcing, enrichment, screening and routing run as agents on the platform, a dashboard reflects live state instead of manual updates, and humans step in only at the high-leverage decision points. Recruiting runs like a product: automated, observable, scalable. You are customer zero, sitting next to the founder, and what you learn shapes the product.

This is a real engineering seat with an unusually wide brief. One week you are shipping a TypeScript agent that turns inbound candidates into a scored pipeline. The next you are in a design-partner conversation, then wiring their feedback into the platform. You will write code most days, and you will decide what to build more often than you are told.

Who we're looking for

Two profiles, one seat. Level, scope, and compensation flex to fit:

• The experienced engineer (roughly 3 to 15 years) who has shipped and maintained production software and can't wait to leave the comfortable yet limiting environment of a big enterprise. Someone who wants to live and work fast to unlock more of their future potential now, and to learn what it takes to do the hard things and what founding a startup is like.
• The ambitious new graduate who already builds with AI daily, has a public trail of shipped projects, and would rather compress ten years of learning into two at a startup than wait in line somewhere large.

You probably

• Use AI coding agents (Claude Code, Cursor, Codex or similar) as a force multiplier, and can read, debug and own every line they produce.
• Are fluent in TypeScript or JavaScript, comfortable with APIs, data pipelines and LLM tooling, and can pick up whatever the problem needs.
• Have shipped real things end to end: an automation, an internal tool, an agent, a product. Side projects count if they ran and someone used them.
• Prefer building the system that fixes a process to managing the process.
• Are comfortable with ambiguity and don't wait to be told what to build next.
• Write clearly and don't let things fall through the cracks.

What you'll build

Agents on our platform: TypeScript running in sandboxed V8 isolates, inference on our own GPU clusters using open models, state in content-addressed storage, signed lineage on every operation. You will use the same runtime, SDKs and CLI our customers use, and you will publish the agents you build into the catalog they draw from.

What the first 90 days look like

• Day one: ship your first agent on the platform and watch its lineage log fill in.
• Month one: you are in the room with design partners. You hear the real problems first-hand, and what you hear becomes platform work.
• Month three: you architect custom solutions on our platform for real customers, end to end: scope the problem, design the agents, ship, and own the outcome.

Compensation

Calibrated to your profile and level, and deliberately upside-oriented: a competitive salary plus meaningful early-team equity, because we expect to go big fast and early team members should share in it. We move fast when there is alignment.

This role is NOT for

• Candidates who want an administrative or coordination role.
• Candidates who need a static job description.
• Candidates who are AI-curious in theory but have nothing shipped.
• Candidates who need heavy structure or oversight to move.

How we hire

• Intro call to align on expectations and fit.
• Founder conversation on vision, mindset and how you think.
• References.
• Paid build trial (about five days): a real problem, our platform, your code. Runs in parallel with the offer conversation.