Ishan Anand

This talk breaks down how Large Language Models (LLMs) like GPT-2 work, demystifying them for web developers by implementing a core GPT-2 model in approximately 600 lines of vanilla JavaScript. The presentation emphasizes understanding the underlying mechanics rather than requiring deep ML expertise, using analogies and code walkthroughs to make complex concepts accessible. The goal is to transform the perception of LLMs from magic to understandable machinery.