Beyond the Chat Box: Context Engineering for Reusable AI Workflows

Whether your lab, your team, or your advisor expects it, AI is becoming part of how graduate work gets done. Yet right now, using ChatGPT or Claude often feels like managing an assistant with amnesia. Every time you tackle a problem set, review a paper, draft an update, or prep slides, you start from zero. You end up retyping the same background context, re-explaining your formatting preferences, and fixing the same avoidable mistakes. Often, managing the prompt takes longer than doing the work yourself.

This hands-on studio fixes that by tackling the two biggest headaches of working with AI:

• Claude Skills eliminate the headache of constantly repeating yourself. Teach Claude your standards and procedures once, and watch it get your workflow right on the first try without copy-pasting old prompts or endless corrections.
• Scheduled Tasks eliminate the headache of doing routine work that can be easily automated. Instead of letting recurring tasks pile up on your to-do list, set Claude to run your workflows automatically in the background.

By the end of the studio, you won't just have just a better understanding of context engineering; you will leave with a working AI setup that knows your standards, runs on your schedule, and takes routine busywork off your plate for good.

Before you come, request your free Stanford Claude for Education account through University IT, install the Claude desktop app, and bring a laptop. No coding experience is needed. Snacks will be provided.

This Engineering Grad Studio is presented as part of the STEM Learning Program at the Center for Teaching and Learning and the Student Success and Engagement Team at the School of Engineering. For more information about the program and upcoming events, visit https://ctl.stanford.edu/about-ctl/events.