Build the trust layer for autonomous work — with people as good as you.
Lucidrail is the governed control plane for AI agents: you set a goal, a team of agents plans, delegates and ships the work, and you stay in command. It is a hard, multidisciplinary problem, and we believe it is solved by a small number of exceptional people. Our job is to build an environment where they can do the best work of their lives.
What we believe
Leadership is service, not control. World-class builders are always in demand, so we don't stand above talented people and dictate outcomes — we build the road, remove the obstacles, and create the conditions where extraordinary people can do their best work. We optimize for talent density over headcount, autonomy over process, builders over spectators, long-term value over short-term optics, clear thinking over politics, and taste and judgment over credentials alone.
Who we hire
- Intelligence — clear thinking that gets to the heart of a problem fast.
- Resourcefulness — finding a way through, with or without a map.
- Speed of learning — picking up new domains quickly.
- Taste and judgment — knowing what good looks like, and what to leave out.
- Ownership — treating the outcome as yours, end to end.
- Systems thinking — seeing how the parts connect.
- Execution under ambiguity — moving while the problem is still undefined.
- Obsession with quality — caring deeply about craft.
What you'll work on
You'll work on the Lucidrail control plane itself — agent orchestration, governance and trust (autonomy levels, approvals, the audit trail), security (the credential broker, isolation and agent attestation), shared memory, multi-model orchestration across Claude, Codex, Gemini, Grok and more, and developer tools and local-compute hardware. The problems are multidisciplinary and frequently undefined, and the people who thrive here move toward complexity, not away from it.
What we offer
- Meaningful ownership
- High trust and high autonomy
- Clear mission alignment
- Access to difficult, important problems
- Minimal bureaucracy and strong operational support
- Real respect for builders and thinkers
Frequently asked questions
What do you look for in candidates?
We care less about pedigree and more about intelligence, resourcefulness, speed of learning, taste and judgment, ownership, systems thinking, the ability to execute under ambiguity, and an obsession with quality.
What kind of problems will I work on?
The Lucidrail control plane itself: agent orchestration, governance and trust, security and attestation, shared memory, multi-model orchestration, and developer tools and local-compute hardware — multidisciplinary, often-undefined problems.
How do I apply?
Tell us what you're great at and what you'd want to build, through the form on the careers page or by email at careers@lucidrail.ai.