Lead – Agent Coding and Acceleration

At Humanoid, we are working from a simple hypothesis: AI is rapidly transforming engineering. The bottlenecks are shifting from writing code to how efficiently AI-authored changes can be verified.

We’re looking for a Technical Lead to build the platform that lets our engineers move faster with coding agents while protecting our code, infrastructure, and robots. You’ll give developers and agents the environments, context, and tools to work effectively, and build verification and security into the path from task to healthy deployment. This means rethinking how engineering work gets done, not just adding agents to existing processes and establishing the access controls, isolation, and monitoring needed to guard against mistakes, misuse, and new security threats.

This is a hands-on leadership role and the keystone hire for a new team. You’ll build the critical systems yourself, recruit and lead engineers, and work across Developer Tools, Security, Safety, Simulation, and robot software. Success means increasing engineering output while reducing cost and human review per accepted change, without increasing defects or safety risk. What you'll do:

Own the path from an agent task to a healthy deployed change, including acceptance criteria, execution, verification, merge, rollout, monitoring, and rollback.

Build secure, reproducible agent runtimes with task-scoped identity, short-lived credentials, bounded access, cost controls, and re-playable execution records. Proactively detect unauthorized agent activity across our compute clusters, whether internal or external.

Build the interfaces developers use to launch, monitor, and control agents, alongside tooling for agents to collaborate through shared memory, coordinated work, and appropriately authorized access to company resources.

Reduce the tokens and retries required for accepted work through better context assembly, model and tool routing, caching, task decomposition, and feedback.

Give agents fast access to builds, tests, simulation, replay, and hardware-in-the-loop, with acceptance criteria they cannot alter.

Work with engineering teams to make the codebase easier to navigate and verify through clear interfaces, strong invariants, reproducible environments, fast local checks, and well-documented implementation paths.

Recruit and lead the team while remaining hands-on. Measure success through cost per accepted change, human review time, escaped defects, and time from push to healthy on the robot.

Experience building and operating developer tooling or infrastructure used across engineering teams, such as build systems, CI/CD, test infrastructure, deployment pipelines, or internal platforms.

Strong software engineering skills and comfort working in large codebases or monorepos.

Practical experience with Linux, containers, and cloud infrastructure, plus working knowledge of Kubernetes

Hands-on experience using and integrating coding agents into real engineering workflows. You understand how context, tools, and evals affect reliability, latency, and cost, and can talk concretely about what worked and what didn’t.

Sound security judgment around workload isolation, identity, permissions, and credentials. You can design appropriate boundaries for agents accessing code, infrastructure, and company resources.

Experience building reliable automated tests or verification systems and diagnosing failures across code, tooling, and infrastructure. You have good judgment about what automated checks establish and where human review still matters.

Experience leading engineers while remaining hands-on. You can set technical direction, recruit a strong team, and turn an ambitious goal into systems that engineers use and trust.