Cantina Labs
Machine Learning Engineer, Ops
Europe +1 · Remote · Full-time · Posted Jul 27, 2026
- Location
- Europe
United States
About the job
Own and productionize the GPU inference stack for real-time streaming TTS and ASR models, optimizing performance, cost, and accuracy.
What you'll do
- Build and optimize streaming inference engines and multi-stage pipelines, including latency and throughput improvements.
- Deploy and operate GPU inference services on Kubernetes; build CI/CD and reproducible model releases.
- Maintain accuracy regression tests and own observability and performance benchmarking; partner with research to productionize models.
What they ask for
- Hands-on experience deploying, scaling, and optimizing ML models in production.
- GPU performance optimization and working knowledge of the inference stack, from CUDA kernels to serving frameworks.
- Production experience with Kubernetes, CI/CD, and cloud infrastructure.
- Strong Python and PyTorch; rigorous performance benchmarking approach.
Nice to have
- Experience with speech or audio models (TTS, ASR, voice conversion).
- Experience with inference frameworks such as vLLM (vLLM-Omni), SGLang, TensorRT-LLM, TensorRT or Triton Inference Server.
Tools
CUDA, Kubernetes, Python, PyTorch, Nsight Systems, PyTorch Profiler, vLLM (vLLM-Omni), SGLang
AI focus
Productionize and optimize real-time streaming text-to-speech and speech recognition models, partnering with research to bring new models into production.
Model-written summary · the original posting is authoritative