Location: San Francisco or Croatia
Type: Full-time
Reports to: CTO
About Daytona
Daytona is the infrastructure layer for AI agents. We provide secure, isolated, instantly available sandboxes where agents run code, use computers, and train. Our platform powers agent workloads, reinforcement learning, and developer environments for some of the largest AI companies in the world. We closed our Series A earlier this year and have already grown 10x since. We are a small, senior team moving extremely fast.
The role
You will lead the pod building Daytona's Windows sandboxes: on-demand, isolated Windows VMs built for AI agents that need to use real desktop software. Computer-use agents, legacy enterprise apps, Office automation, QA at scale. Nobody has made Windows feel serverless. That's the job.
This is an engineering role first. You will build alongside the team most of the week and own the technical direction, and you will also manage the pod: 2 senior engineers with a dedicated PM. Management is part of the job, not the whole job.
What You'll Do
- Own architecture and delivery of the Windows sandbox runtime: Windows guests on our KVM-based stack, boot and login time optimization, snapshot/pause/fork for Windows VMs
- Drive Windows-specific hard problems: image build pipelines, sysprep and provisioning automation, driver and virtio integration, display and input streaming for computer-use agents
- Own the licensing-compliant architecture for Windows in a multi-tenant cloud (SPLA and BYOL realities) together with legal and finance
- Own reliability, cost per sandbox, and density on Windows hosts
- Hire and lead the pod; partner with your PM on roadmap and enterprise commitments
What We're Looking For
- 8+ years in systems or infrastructure engineering, 2+ years leading engineers
- Deep experience running Windows at scale in non-Microsoft-stack environments: Windows guests on KVM/QEMU, virtio drivers, image automation
- Strong Windows internals knowledge: boot process, services, registry, WinRM/PowerShell automation, unattended provisioning