Member Of Engineering

Summary

Join poolside, a company building AI to drive economically valuable work and scientific progress, as a Performance Engineer. You will accelerate high-level primitives used to train Large Language Models (LLMs) and optimize distributed communication. Your mission is to ensure cutting-edge performance of LLM pre-training and fine-tuning on large GPU clusters. This role requires expertise in CUDA, C++, Cython, and Triton programming, along with a deep understanding of GPU hardware and architecture. Poolside offers a fully remote-first work environment with flexible hours and generous benefits.

Requirements

• Engineering background
• Expert understanding of GPU hardware/architecture
• Strong C/C++ programming skills
• Fine-grained knowledge of CUDA programming
• Strong algorithmic skills
• System programming on Linux experience

Responsibilities

• Profile CPU and CUDA code at several abstraction levels
• Debug and profile distributed applications
• Troubleshoot undocumented CUDA internals
• Hack the NCCL library used for GPU communication
• Tune vanilla CUDA, Triton, CUTLASS kernels for the latest NVIDIA GPUs
• Hack PyTorch internals

Preferred Qualifications

• Knowledge of CPython internals and experience of native extension development
• Knowledge of AWS EFA internals
• Compiler development background

Benefits

• Fully remote work & flexible hours
• 37 days/year of vacation & holidays
• Health insurance allowance for you and dependents
• Company-provided equipment
• Wellbeing, always-be-learning and home office allowances
• Frequent team get togethers
• Great diverse & inclusive people-first culture