Systems Engineer in San Francisco

Dedalus Labs is hiring a Systems Engineer with depth in OS and kernel internals, virtualization, file systems or distributed infrastructure. You will work at the lowest layers of the stack (kernel, hypervisor, storage, networking) and own entire subsystems end to end on an 8-person team where your work ships to production immediately. New grads with an exceptional body of work and experienced engineers (3-6+ years) are both welcome; the CTO invests heavily in growing engineers who ramp fast., * Build core compute and runtime primitives for long-running, stateful AI agents, including virtualization, sandboxing and isolation for secure multi-tenant workloads.

• Own subsystems end to end (storage, scheduling, networking, virtualization) and ship them to production.
• Design persistent state, snapshotting and recovery mechanisms for always-on agent workloads with the CTO.
• Drive sandbox startup latency below 50ms and optimize every syscall and scheduling decision in the critical path.
• Build distributed storage layers, schedulers and networking infrastructure that scale across data centers.
• Profile, debug and benchmark systems under real-world scale, latency, correctness and reliability constraints.
• Contribute to the roadmap for on-prem GPU clusters and model training infrastructure., * Virtualization, hypervisors, KVM, Firecracker or container runtimes
• Distributed storage, replication, consensus or consistency models
• GPU infrastructure or AI/ML systems (training clusters, on-prem GPU setup)

• Proficiency in Rust, Go, C or C++ for low-level systems programming (Rust )
• Depth in OS/kernel internals, file systems, networking or low-level memory management
• A verifiable body of work: open-source contributions (e.g. kernel patches), systems papers, technical blogs or substantial GitHub activity
• A bachelor’s degree or equivalent rigorous program
• Evidence of exceptional drive (competitive programming, hackathon wins, competitive sports, technical writing or talks)
• Ability to work on-site in San Francisco 5 days a week

Dedalus Labs is building the compute substrate for an agent- economy. Its flagship product, Dedalus Machines, gives AI agents fast, persistent virtual computers that boot in under 50 milliseconds and run long-lived, stateful workloads that never sleep. The platform also includes an MCP gateway and Agents SDK that unify the fragmented agent ecosystem behind a single API.

Founded in 2025 and a Y Combinator (S25) company, Dedalus Labs has about 8 people in San Francisco and has raised $11M led by Kindred Ventures and co-led by Saga Ventures.