Pinned Loading

1. llms-from-scratch-cudallms-from-scratch-cudaPublic

   A collection of LLMs implemented from scratch in CUDA/C++ — every kernel hand-written, zero frameworks.

   Cuda

2. RFX-FuseRFX-FusePublic

   Breiman and Cutler's Random Forests Compressed as a Unified Learning and Similarity Engine. Extended with native explainable similarity. Scales to 25M+ with GPU acceleration.

   C++ 9

3. speakingspeakingPublic

   Technical presentations on production ML systems, LLMs, and enterprise AI.

4. Text-2-3D-printText-2-3D-printPublic

   Text-to-3D-print pipeline: fine-tuned SDXL LoRA generates images, InstantMesh converts them to printable meshes.

   Jupyter Notebook