Senior Lead ML Platform Engineer, Training and Inference

Sourced from simplify.jobs

Paramount Global's Applied Machine Learning Group is hiring a Senior Lead ML Platform Engineer to own the technical direction for its training and inference infrastructure. The role leads adoption of Ray/AnyScale for distributed training and architects a high-performance Kubernetes-based inference layer serving billions of requests per day. You will optimize GPU utilization and cost while setting organizational standards for ML deployment, monitoring, and reliability.

• Own the long-term architecture for training and inference domains, scaling the platform 10x over 1–3 years
• Lead implementation and optimization of Ray/AnyScale as a unified distributed compute layer
• Design and maintain K8s-based inference servers (Triton, TorchServe, vLLM) optimized for GPU memory and throughput
• Optimize GPU instance trade-offs (A100, H100, T4) for cost, availability, and performance
• Establish reusable patterns for CI/CD, model versioning, and canary deployments across teams
• Define and enforce SLIs/SLOs for platform reliability
• Mentor senior engineers across ML Platform and Applied ML pods

• 6–8+ years in ML infrastructure, platform engineering, or high-scale backend engineering
• Extensive experience with Kubernetes and large-scale ML serving frameworks
• Strong knowledge of GPU architecture, CUDA, and hardware acceleration
• Proven track record owning technical direction for a major domain across multiple teams

• Infra-as-code with Terraform or Pulumi and automated MLOps pipelines
• Deep expertise with Ray/AnyScale or similar distributed compute frameworks
• ML observability tools (Prometheus, Grafana, Weights & Biases, MLflow)
• Multi-cloud or hybrid-cloud ML environments
• Python and C++ for performance-critical systems