AI Summer

Summary repository for AI Summer 2026. Introduction to generative AI, with practical applications for inferencing and training May 5-30, 2026

Presented by Vanderbilt Data Science Institute data scientists:

• Dr. Jesse Spencer-Smith, Chief Data Scientist
• Umang Chaudhry, Senior Data Scientist
• Myranda Shirk, Senior Data Scientist

Overview

The objective of these workshops is to develop foundational skills in understanding, inferencing and training generative AI models and other transformer models.

Getting Ready for AI Summer

Getting Accounts

You’ll want to use the most advanced AI chat model that you can get access to. To get a headstart, please create an account with https://claude.ai/ . We will be spending a signficant time using Claude, Claude Cowork and Claude Code. Please sign up for a paid Claude account to gain access to these tools.

Decide on a Project (recommended)

Think about any data you might want to bring to the workshop. Also begin thinking about any projects you might want to accomplish during our month. We’ll have office hours for you to work with us to get your first project off the ground!

Workshop Schedule and Recordings

Session will run live from 11am-1pm, with an office hour from 1pm to 2pm (all times Central).

All session recordings are posted to the AI Summer 2026 YouTube playlist.

Week 1

Day 1 — Tuesday, May 5, 2026

Topics covered:

• Working with secure data: Vanderbilt’s data classification levels (1–4) and the AI tool that fits each one
• Opting out of training, ChatGPT EDU, Vanderbilt’s Amplify, and where Claude / Cowork / Code fall in the framework
• Working locally: open-source models, quantization, distillation, and PEFT
• Demo: LM Studio for fully-local model use
• Tour of Claude Chat vs. Claude Cowork vs. Claude Code (desktop and CLI)
• What separates a chat from an agent: planning, task files, sub-agents, evaluation
• How LLMs actually work: tokens, embeddings, attention, MLP, temperature
• Context as the central problem: tools, MCPs, RAG, and context rot
• Skills and the agent harness (“Agent 2.0”); progressive disclosure
• Live build: co-creating a “sensitive field detector” skill in Claude Cowork

Resources:

• Lecture notes
• Recording on YouTube

Day 2 — Thursday, May 7, 2026

Topics covered:

• Recap: chat vs. agent (replay), and what “Agent 2.0” means
• Walkthroughs of two real skills: the sensitive-field detector and the lecture-recording-to-lecture-notes skill that produced these notes