Hi, I'm Sam.

I'm a systems developer in Vancouver interested in GPU computing, AI, and mathematics.

I'm currently working on FindTensor, an experimental compiler and runtime project for LLM inference.

I enjoy understanding systems close to the hardware and care about careful benchmarking, reproducible results, and correctness.

A few things I've worked on

• FindTensor — an ongoing exploration of compiler and runtime techniques for LLM inference.
• Lean formalizations — formalization work in Lean 4 and mathlib around several open problems in combinatorics.
• GPU architecture research — worked with Professor Tor Aamodt at UBC ECE on modeling newer GPU architectures in GPGPU-Sim.

Tools I've worked with

Python, Rust, C++, CUDA, PTX, NCCL, Lean 4, PyTorch, vLLM, Bazel, and Linux.

Contact

findtensor.com · LinkedIn · Email