CMSE 492 (aka CMSE 101) • Fall 2026 Prof. Danny Caballero
For four weeks we built DTPA one piece at a time: Data (Week 1), Tools (Week 2), Practices (Week 3), Actions (Week 4).
Starting today, you use all four at once, and you find the case.
Your group picks one broad domain, then narrows to one specific, documented case inside it.
Domain is not a case. "AI in healthcare" is not a case; a specific hospital's use of a named documentation tool is.
Your case needs: a named tool, a named organization, and a documented deployment (something that happened, with a date, from a citable source).
Two things happen in parallel:
Same template family as your weekly Case Studies, split into two parts:
Use Section 0 to dump links, notes, and half-finished thoughts as you go — nothing there is graded, but nothing gets lost either.
Posted on D2L and linked from the Week 5/6 schedule page.
Do not stop at the first case you find, and do not default to the domains materials we posted.
Those readings are the most visible stories in each domain — that's exactly why they're a place to start, not a place to finish.
An independent research option on the energy, water, and political costs of AI computing.
Every topic must address two required lenses:
Choose at least one: AI policy & power · energy/water use case studies · open vs. closed systems · energy sources · edge computing · tokens as an efficiency metric · labor behind the systems.
Format (your choice): ~3,000-word essay · 10-minute video + slides · scientific poster — must include a table of energy/cost calculations and your proposal.
Milestones:
Starting Week 11, groups can opt to lead half of a Monday or Wednesday class, walking the room through one of your group's three Full Cases (I, II, or III) using DTPA.
Slots are one group per half-class, first-come first-served — email me to claim one.
You have the rest of class to begin. By the time you leave today, your group should have:
You have all week. The goal today is momentum, not a finished case.