Research Engineer, ML Systems (All Industry Levels)
Core
Designing and optimizing high-performance ML training and inference systems for GPU clusters to serve LLMs at scale.
Role type
Research Engineer, ML Systems
Builds
GPU clusters, distributed RLHF stacks, multimodal model training/inference systems
Domain
Artificial Intelligence / Machine Learning Systems
Deliverable
production ML models
Required skills
PyTorch, distributed machine learning, reinforcement learning, transformers, Triton kernel development, CUDA programming, LLM training and distillation
Preferred skills
DeepSpeed, Megatron, vLLM, FlashAttention, Kubernetes, Docker, cloud orchestration, academic publications
Responsibilities
Write efficient Triton kernels for specific models and hardware, develop prefix-aware routing algorithms, train and distill LLMs, build distributed RLHF stacks, develop systems for multimodal model training and inference
Seniority
All Industry Levels (PhD or equivalent research experience required)