We're a small, intense team building the infrastructure layer that makes AI inference fast, reliable, and deployable anywhere. If you want leverage over headcount, this is the place.
We don't have a ping-pong table. We have hard problems, real equity, and the autonomy to do your best work.
RL-driven inference routing, distributed scheduling across heterogeneous clusters, and compiler-level optimizations. This is systems work at the frontier.
Under 30 people. Every engineer ships to production. No layers of management, no approval chains. Your work matters on day one.
Distributed across US and EU time zones. Async by default. We care about output, not hours logged. Optional co-working in San Francisco.
Meaningful ownership with a real path to liquidity. Top-of-market cash comp, full benefits, and equipment budget. We invest in the people who invest in us.
We hire slowly and deliberately. Every role on this list is one we need filled yesterday.
Design and optimize the reinforcement learning policies that drive real-time inference routing across our global mesh.
Own the core orchestration layer: scheduler, placement engine, and the control plane that ties it all together.
Deploy and operate SpinDynamics in air-gapped, on-prem, and hybrid environments for our most demanding enterprise customers.
Build the observability dashboard, deployment console, and developer experience that makes complex infrastructure feel simple.
Work on model graph optimization, kernel fusion, and cross-backend compilation targeting heterogeneous accelerator topologies.
Be the bridge between our engineering team and the developer community. Write docs, build demos, and shape how the world discovers SpinDynamics.
Design deployment architectures for enterprise customers, translate business requirements into technical solutions, and drive adoption post-sale.
We're always looking for exceptional people. If you've built systems that others depend on and want to work at the intersection of ML and infrastructure, send us a note.