Senior MLOps Engineer

About the role

What you'll do

• Deploy and scale self-hosted open-weight models from approximately 7B to 370B or more parameters using serving engines such as vLLM, Triton, or TGI.
• Choose serving strategies including continuous batching, tensor or pipeline parallelism, and quantization to meet latency and cost targets.
• Operate a multi-provider gateway spanning self-hosted models and external APIs, with token accounting, budget controls, and cost-aware routing.
• Build and maintain fine-tuning, evaluation, versioning, and continuous delivery pipelines using MLflow, SageMaker, or Kubeflow.
• Support distributed training with DeepSpeed, FSDP, or Accelerate.
• Own monitoring, logging, alerting, incident response, and safe rollback for production ML services.

What we're looking for

• 5+ years of experience in MLOps, ML infrastructure, or ML engineering, with ownership of end-to-end production model lifecycles.
• Hands-on experience serving large models at scale, including vLLM, Triton, or TGI, quantization, and GPU optimization.
• Strong experience with Kubernetes, Docker, and Terraform.