How does software engineering change when AI is part of the product we're creating?
Current practice suggests a need for structured experimentation, tolerance for uncertainty, deep qualitative research and evaluation, wrangling with questions of meaning and knowledge, advanced statistics and mathematics, and more.
How might we expand our efforts to build these skills?
CDSC faculty member Kaylea Champion recently received the very welcome news that, together with her Co-PI Jeffrey Kim, she was awarded a grant from the UW’s internal AI-SEED program to explore this key topic in empirical software engineering and computer-supported cooperative work!
The official press release from UW — including many other exciting awards supporting innovation in teaching with / about AI — is here.
Grant Title: “Rethinking engineering workflows: Building AI-powered software.”
Grant abstract: “This project focuses on teaching students the impacts to a software engineering workflow — including planning and estimation, requirements analysis, design, development, testing, deployment and maintenance — when AI-powered features are part of the product. Although substantial attention has been paid to AI as a code generating tool and to the technical skills involved in building AI models, impacts of AI features on overall engineering workflow is relatively less explored. All computing students should have an awareness of these process impacts, regardless of their specialties and goals. We propose to develop and evaluate two interventions: a no-code hands-on simulation of evaluating and fine-tuning a model, and a series of enhancements and extensions for learning modules focused on different stages of a development project. We propose to develop a preliminary version of these interventions, then to evaluate and refine them through a series of co-design workshops, and finally to publish our results.”
Discover more from Community Data Science Collective
Subscribe to get the latest posts sent to your email.