Founding EngineerAI

We're building AI systems for commercial HVAC service — an industry that runs on tribal knowledge, paper trails, and senior technicians who can't be cloned..

We're not a deck looking for a customer. We have a key enterprise customer, a live operating company we own as our R&D environment, and working integrations already in production. The operating company — a commercial HVAC service business in Seattle — is our sandbox: real technicians, real equipment, real decisions getting made every day. Whatever you build, you can deploy it Monday morning to people who'll tell you exactly what's helpful vs not.

The founding team has previously built and exited multiple startups, paired with a backend and data team that's deeply experienced in production systems. You'd be joining people who have done this before.

The Role

You'd own the AI layer end to end — the agents, the orchestration, the workflows that wrap them. You're the senior technical voice for everything model-facing: how agents are designed, how they cooperate, how we measure whether they're actually working, and how we iterate without regressing.

You'll work alongside a team with deep domain expertise and direct customer relationships. They know the industry, the workflows, and the people we're building for. Your job is to translate that into AI systems that actually work in the field.

In your first six months, you'd likely:

• Design and ship the next generation of our field-facing AI assistant — voice-first, decision-trace-capturing, embedded in the technician's day
• Build the agent and orchestration framework that lets us compose multi-step workflows across diagnostics, planning, and field execution
• Stand up the eval, retrieval, and observability infrastructure that lets us iterate on agent behavior without breaking what works
• Productionize agentic workflows that turn unstructured field activity into structured operating intelligence

Backend, data infrastructure, and integrations are owned by an experienced team that's been through this before. Your job is the intelligence on top.

What we're looking for

A builder who wants to own it

• Senior or Staff-level experience shipping production software as a hands-on builder — not a manager, not a reviewer, the person who actually makes the thing
• You want decisions, not direction. You're looking for a role where the technical calls on the AI layer are yours to make, defend, and live with
• Strong product judgment. You can take a vague problem, find the high-leverage version of it, and ship something pragmatic
• Comfortable with the founding-team reality: high autonomy, minimal oversight, no one to hand the hard problems to

AI engineering

• Demonstrated taste applying LLM APIs, retrieval systems, agent frameworks, and orchestration patterns to real workflows — production, not toy
• You've shipped multi-step agentic systems that interact with external tools and stay working. You know when to use one model call vs. an agent vs. a deterministic workflow, and you can defend the choice
• You know how to build the supporting infrastructure that makes AI reliable — evals, observability, prompt versioning, retrieval quality measurement — and you know when in a product's life it's time to invest in it
• You understand the limits of the model and own correctness and security. AI accelerates you; it doesn't absolve you
• You write most of your code with AI tools like Claude Code, or Codex, and already manage multiple sessions in parallel

Foundations

• Strong systems design instincts. You've built scalable services and integrated against messy real-world systems, even if that's not your day-to-day going forward
• You've shipped production software end-to-end and understand what it takes to keep it running

You ship fast, then iterate

• You believe the first version exists to prove the idea, not to scale. You ship to learn, you let the customer tell you what matters, and then you harden what's earned its place
• You can take an idea from 0→1 faster than most people can write a PRD
• You're relentlessly pragmatic. You make the smart tradeoff to get value out the door, then iterate once you know what matters
• You're energized — not depleted — by messy, high-constraint environments: legacy codebases, enterprise systems with incomplete documentation, integrations that fight back

Pay: From $150,000.00 per year

Work Location: Remote