Priya Aryav

A small autograd engine and neural net library.

Created a small byte pair encoder to encode text to tokens and vice versa.

A tiny pre-trained implementation of GPT, GPT-2, Qwen2.5 and other architectures

WIP - LLM training and inference in simple, raw C/CUDA, and Triton

WIP - A teeny tiny ChatGPT

Implementing ML Algorithms, Model Architectures, Pre-training, Post-training, RL

Explained everything about how to Build GPT from scratch in PyTorch.

Detailed explanation about attention mechanism used in transformer architecture.

All about matrices and their operations, math, and intuition used to build AI.

Built a minimal tokenizer to convert words to tokens used in LLMs.