In this talk, we will explore how Large Language Models (LLMs) can autonomously perform tasks traditionally handled by data scientists. Using live coding, we will demonstrate how LLMs can explore a dataset, generate hypotheses, write and test code, and fix issues as they arise.

We'll also cover how LLMs can test statistical significance, draw charts, and interpret results-capturing the essence of what a data scientist does. Additionally, we'll discuss the evolving role of human data scientists in a world where LLMs can handle so much of the data science workflow, and examine where human expertise will still be essential in the process.