DS-4XX · Harnessing Language Models for Data Science
A proposed advanced course on using language models in data-science work: prompt design, RAG, AI-assisted analysis, code generation, validation, attribution, and responsible workflow design.
This proposed course is about using large language models as part of a serious data-science workflow without letting them blur the line between support and evidence.
What the course would cover
Students would build AI-assisted workflows for data exploration, reporting, documentation, code generation, and retrieval-augmented analysis. The technical work would be paired with verification habits: checking outputs, documenting where AI helped, protecting sensitive data, and knowing when automation makes a claim weaker rather than stronger.
Want to get involved?
If this course sounds like something you want to take before it exists as a regular offering, talk to me. A directed-study version is the right shape for students who want to help pressure-test the syllabus, build examples, or adapt the ideas to a real data-science project.