We're a small, focused team solving hard problems in AI infrastructure. No meetings that could be emails. No sprints — just deep work on systems that run at the edge of physics.
Compiler optimization, kernel development, GPU scheduling. Real systems programming, not YAML configuration.
Work from anywhere. We optimize for output, not hours online. Deep focus time is sacred.
Early-stage equity with transparent valuation. You're building the foundation — you should own a piece of it.
CUDA, NVLink topology optimization, custom kernel development for inference workloads.
Build compilation pipelines for ML workloads. Kernel fusion, operator scheduling, memory optimization.
Design fault-tolerant agent orchestration networks. gRPC, NATS, CRDTs, consensus protocols.
Bare-metal fleet management, observability at scale, thermal monitoring, and incident response automation.
Build the control plane UI for GPU cluster management. Real-time dashboards, topology visualization.
No trick questions. No whiteboard algorithms. We want to see how you think about real systems.
30 minutes. Tell us what you're excited about. We'll share what we're building.
Walk us through a system you've built. We'll ask questions about tradeoffs and failure modes.
A focused 1-week project related to actual work. Paid at your full rate. No free labor.
Clear comp breakdown: salary, equity, benefits. No negotiation theater.
We're always open to exceptional engineers. Tell us what you'd build.
Send Open Application