Research Engineer, GPU Performance
Core
Develop low-level code to maximize GPU capacity for training and serving frontier large language models, optimizing memory layout, communication, and model architecture for efficient inference.
Role type
Senior IC research engineer (GPU performance & distributed systems)
Builds
High-performance training and inference pipelines for frontier LLMs on large-scale GPU clusters
Domain
Generative AI, High-performance computing, Distributed systems
Deliverable
production ML models
Required skills
Custom low-level kernel development, Distributed GPU cluster infrastructure, Generative AI model architecture, Performance profiling and benchmarking, CI/CD integration, Software design
Preferred skills
Experience with LLM fine-tuning, Tooling/DevOps/infra engineering
Technologies
High-end GPUs, MLOps frameworks, Large-scale GPU clusters
Responsibilities
Write custom kernels and optimize memory/communication for GPU utilization, Rethink model architectures for efficient inference, Integrate optimizations into MLOps frameworks, Profile and benchmark large-scale training/inference, Model LLM execution on various GPUs for performance estimates
Seniority
Senior, hands-on IC