Member of Technical Staff - Inference
Hyperbolic · San Francisco, CA
Infrastructureonsite
At a glance
Highlights
- Build inference on Forge so customers consume tokens without managing GPUs.
- Own end-to-end model serving across worldwide clusters as primary seat.
- Give NeoCloud partners a full-stack path to their own token-factory offering.
Ideal candidate
A self-directed generalist who owns products end to end and thrives on hard infrastructure problems. You care deeply about making AI compute accessible and enjoy shaping ambiguous areas with minimal direction.
Requirements
- Strong general inference background with broad command from request to token
- Deep Kubernetes experience operating clusters in production
- Solid grasp of inference performance concepts like TTFT, disaggregated inference, speculative decoding, KV cache
- Familiarity with modern inference frameworks and serving engines with judgment to select for workload
- Working knowledge of NVIDIA Dynamo in distributed serving architecture
- Experience setting up monitoring, gateways, and endpoints for production inference
- Proven ability to build product end to end from nothing to real traffic
- Strong self-initiative operating as primary owner with minimal direction
Lab profile
About Hyperbolic
Unifying fragmented GPU capacity into one liquid platform to power AI compute.
What they build
- Unified GPU compute platform
- Marketplace to access, sell and resell GPU capacity
- GPU management and orchestration tools
Known for
- Heterogeneous GPU aggregation
- Global AI compute orchestration
- Neocloud infrastructure enablement
Posted · Last seen