Daytona
Team Lead, GPU Sandboxes
San Francisco · Full-time · Posted Oct 30, 2026
- Experience
- 8+ years in systems or infrastructure engineering; 2+ years leading engineers who ship.
About the job
Lead a four-engineer pod while building and taking Daytona’s serverless GPU sandbox runtime into production, focusing on GPU passthrough, fleet efficiency, reliability, and security.
What you'll do
- Own architecture and delivery of the GPU sandbox runtime, including the transition to full VFIO passthrough.
- Improve fleet efficiency, including bin-packing, cold-start latency, GPU utilization, and idle cost.
- Manage and grow the engineering pod; partner with its PM on roadmap, pricing inputs, and customer commitments.
What they ask for
- 8+ years in systems or infrastructure engineering and 2+ years leading engineers who ship.
- Deep hands-on experience with Linux virtualization: KVM, QEMU, libvirt, VFIO/PCIe passthrough, and IOMMU.
- Production experience running GPU infrastructure at fleet scale.
- Strong systems programming background in Go, Rust, or C/C++; able to own outcomes end to end, including on-call.
Nice to have
- Experience with NVIDIA data center GPUs (H100/A100), NVML, DCGM, or MIG tradeoffs.
- Background at a GPU cloud, hyperscaler compute team, or serverless platform; experience serving AI agent or RL training workloads.
Tools
KVM, QEMU, libvirt, VFIO/PCIe passthrough, IOMMU, Go, Rust, C/C++
AI focus
Build GPU sandboxes for agent-driven ML workloads, reinforcement learning training, and inference experimentation.
Model-written summary · the original posting is authoritative