Systems Engineer (AI Agent Infrastructure)

An early-stage infrastructure startup is building long-lived, isolated computing environments where AI agents can run software and keep their state over time. It is hiring a systems engineer to work on the core platform, directly alongside the CTO and a small group of engineers, with plenty of freedom to choose an approach.

Areas of work

Own the control plane: the interfaces and backend services that provision, place and manage the environments, with careful handling of concurrency, persistent state and failure recovery

Improve the virtual machine monitor and runtime, which are written in Rust, including guest memory handling, device emulation, Linux I/O, snapshots and startup behavior

Profile and cut performance costs in request and startup latency, kernel call overhead, memory allocation, scheduler behavior, and network and disk I/O, backing each change with baselines and measurements

Design secure interfaces with explicit trust boundaries, and find security weaknesses with a demonstrated way they could fail or be attacked

Carry projects from an initial experiment to a verified, deployed and operated service

Who fits

Strong Linux internals knowledge and the curiosity to dig beneath an API

Depth in either control-plane services (Go, Rust or similar) or practical Rust plus low-level systems work; spanning both is welcome

A track record of picking things up quickly and delivering, with recent projects you can walk through in detail

Security-minded precision about real versus speculative risk

Clear written and spoken communication, including design notes and code reviews

Responsible use of contemporary AI coding assistants on top of solid fundamentals

Any career stage is acceptable; demonstrated ability matters more than years or pedigree

On-site in the San Francisco Bay Area. Visa transfers and new sponsorships are both considered.