Senior LLM Engineer
Core
Senior LLM Engineer responsible for the full model lifecycle including data, distributed training, post-training, compression, evaluation, and efficient inference for foundational AI models.
Role type
Senior IC machine-learning engineer (LLM)
Builds
Compressed and optimized LLMs, RL environments, and data pipelines for reasoning, coding, and agentic tasks.
Domain
Artificial Intelligence / Quantum Computing
Deliverable
production ML models
Required skills
PyTorch, distributed training (FSDP, DeepSpeed, Megatron-LM), post-training (SFT, DPO, RL), model compression (quantization, pruning, distillation), GPU fundamentals, experimental design, data pipeline engineering
Preferred skills
NVIDIA stack (Megatron, NeMo), RL infrastructure (vLLM, SGLang), post-transformer architectures (SSMs, linear attention), Physical AI, multimodal training, inference on edge hardware, large-scale data processing, publications
Technologies
PyTorch, FSDP, DeepSpeed, Megatron-LM, Slurm, K8s, vLLM, SGLang, GPTQ, AWQ, Triton, CUDA, ONNX Runtime, OpenVINO
Responsibilities
Train and post-train LLMs on multi-node GPU clusters; design compression and healing techniques; build RL environments and reward functions; research and prototype new architectures; build data pipelines for curation and synthetic data; design rigorous evaluations; optimize models for inference on various accelerators; write reports and papers; review code and mentor teammates
Seniority
Senior, hands-on IC
