- United States +1 more
- Full-time · Senior
- 8+ years
Own the end‑to‑end storage architecture for Runware’s AI supercomputer, designing, building, and operating storage systems across US and Europe pods.
Locations (2)
- United States
- United Kingdom
Responsibilities
- Own the end‑to‑end storage architecture for the fleet
- Design and build the model distribution path
- Drive model load latency down
- Select, benchmark and operate the storage systems
- Set and hit reliability and durability targets
Requirements
- Nice to have: GPUDirect Storage
- Nice to have: NCCL
- Nice to have: vLLM
- Nice to have: SGLang
- Nice to have: TensorRT-LLM