Hello, I'm jrhemstad π
- π I'm currently the Engineering Manager for the team that maintains the CUDA C++ Core Libraries at NVIDIA.
- π± Previously, I was deeply involved in the RAPIDS ecosystem, where I was a primary maintainer and designer for the libcudf and RMM libraries.
- π I helped kickstart and maintain the cuCollections project for GPU-acelerated data structures like hash tables.
- π‘ I'm passionate about building high-quality, high-performance CUDA C++ software.
- π― My mission is to make CUDA C++ more delightful for developers.
- Revamping llm.c with the CUDA C++ Core Libraries
- GTC 2024: Mastering CUDA C++: Best Practices with the CUDA C++ Core Libraries
- Using CUDA Stream-Ordered Memory Allocator (Part 1)
- Using CUDA Stream-Ordered Memory Allocator (Part 2)
- Maximizing Performance with Massively Parallel Hash Maps on GPUs
- GTC 2019: RAPIDS CUDA DataFrame Internals for C++ Developers
- π LinkedIn