Paarth Sharma

Beyond Tutorials: Training My Own GPT in JAX

There's a point in every ML engineer's journey where tutorials stop being enough. You follow along, you copy the code, the loss goes down, and you think you understand transformers. Then someone asks