Senior/Lead Research Engineer (Foundation Models)
Core
Build and operate systems for training, evaluating, and running the Thomson LLM family of foundation models for professional workflows.
Role type
Senior/Lead Research Engineer (Foundation Models)
Builds
The Thomson LLM family (custom models for professional workflows)
Domain
AI / Large Language Models / Reinforcement Learning
Deliverable
production ML models
Required skills
Large-scale distributed systems, PyTorch, Megatron-Core, FSDP/DTensor, DeepSpeed, parallelism strategies, LLM inference engines (vLLM, SGLang), GPU profiling, evaluation suite scaling
Preferred skills
RL post-training frameworks (NeMo-RL, verl, OpenRLHF, TRL), evaluation frameworks (Inspect, Harbor, lm-evaluation-harness), distributed orchestration (Ray, Kubernetes, Slurm), open-source contributions (vLLM, Megatron)
Technologies
PyTorch, Megatron-Core, FSDP, DTensor, DeepSpeed, vLLM, SGLang, Ray, Kubernetes, Slurm, AWS, Azure, Google Cloud
Responsibilities
Architect and optimize infrastructure for training agentic Mixture-of-Experts models; build and scale evaluation suites and leaderboards; profile and optimize for cost, latency, and GPU utilization; debug distributed systems and performance issues; set standards for code quality and reproducibility; mentor engineers (Lead level)
Seniority
Senior (hands-on IC) / Lead (architecture, planning, mentorship)