ElevenLabs
Research Engineer - Inference
- United Kingdom
- Remote
- Permanent
Owns the path from research checkpoint to production serving infrastructure, optimising inference latency, throughput and cost with quantization, distillation, KV-cache tuning, batching and custom kernels. Builds high-performance streaming serving systems and tooling so researchers can ship models quickly and safely.
CUDA · Triton · TensorRT +2