Director, Model Inference Engineering
Core
Build and operate Mariner, Thomson Reuters' custom inference platform serving production traffic for the Thomson LLM model family and frontier models.
Role type
Senior IC Director, Model Inference Engineering
Builds
Production model-serving infrastructure (Mariner) for Thomson Reuters' AI products
Domain
AI/ML Infrastructure, Large Language Models (LLMs), Cloud Infrastructure
Deliverable
production ML models
Required skills
Kubernetes, model-serving infrastructure, inference optimization, distributed systems, production reliability, system design, engineering best practices, team leadership
Preferred skills
LLM-serving stacks (vLLM, TensorRT-LLM, Triton, SGLang), GPU-cluster capacity planning, regulated environments (ISO 27001), systems languages (Rust, Go), systems-level optimization
Technologies
Kubernetes, vLLM, TensorRT-LLM, Triton, SGLang, Rust, Go
Responsibilities
Build and operate model-serving infrastructure, implement inference-optimization techniques, improve observability, implement compliance controls (ISO 27001), manage cost vs. latency tradeoffs, collaborate with training/evaluation pods, establish team practices and tooling
Seniority
Senior, hands-on IC with leadership responsibilities