Things I've built, mostly around kernels, compilers, and making models faster.
Custom CUDA kernels and operator fusion for RF-DETR — faster inference out of the box, no retraining or config required.
A minimal PyTorch-like deep learning library focused on custom autograd engines, backpropagation, dynamic computation graphs, and efficient tensor operations written from scratch.
A lightweight, experimental compiler & interpreter framework exploring compiler frontends, AST parsing, intermediate representation (IR), optimization passes, and a custom VM.
Optimized fork of litdata featuring multi-node cluster streaming for deep learning datasets, with specialized caches for CUDA pinned-memory layout.
A curated roadmap of experimental projects spanning CUDA kernels, custom CPU schedulers, microkernels, and lightweight inference engines.