Hello! 👋

My name is Hwanseo Choi. I'm a cuDNN Library Engineer at NVIDIA in Santa Clara, California, where I work on cutting-edge deep learning infrastructure and GPU-accelerated computing.

💼 Experience

• Deep Learning Library Software Engineer @ NVIDIA
  → Added depthwise convolution kernels to cuDNN, contributed to C++/Python frontend for Scaled-Dot Product Attention, expanded testing and reference for multiple variants, and helped integrate cuDNN Scaled-Dot Product Attention into various deep learning frameworks such as PyTorch, JAX, and TransformerEngine.

• Course Assistant @ UIUC
  → Developed autograders, created documentation, held office hours, and graded programming assignments for ECE 220: Introduction to Computer Systems.

🎓 Education

• I received my Bachelor of Science in Computer Engineering from the University of Illinois at Urbana-Champaign. My coursework included topics in parallel programming, artificial intelligence, operating systems, computer architecture, digital systems, and interactive graphics.

🧪 Projects

• MXNet Convolution — CUDA kernel optimization and convolution operator extension; won first place in performance competition
• Dancing Altera Revolution — DDR-style game on Altera NIOS II FPGA using VGA & sound card drivers
• Ray Tracing with BVH — 3D ray tracer in C++ with GLM, 10× speedup using BVH (binary space partitioning)
• Twitch Prediction Tools — Bot using Twitch GraphQL + Riot API to predict match outcomes and farm points
• Live Chess Analyzer — Converts screenshots into FEN, overlays engine evaluations visually in real time

🛠 Skills

• Linux, C/C++, Python, PyTorch, CUDA, cuBLAS, cuDNN, CMake, Docker

🌍 Languages

• English and Korean

📫 Links

• Portfolio
• Blog
• LinkedIn
• GitHub
• X

You can ask me anything (within reason). I'm always interested in discussing deep learning, GPU computing, and other technical topics!