AI-Assisted Development Practices

This unit teaches students to use AI tools throughout the software development lifecycle, including generating code from specifications, refactoring, linting, testing, and maintaining projects across platforms like Copilot, ChatGPT, Sourcegraph, Replit, and Kaggle. Students will practice evaluating and blending AI-generated and human-written code, and learn the fundamentals of transformer models that power these assistive tools.

Learning Objectives

• Analyze AI-generated code against a provided Software Requirements Specification (SRS) to identify functional mismatches, style violations, and potential bugs, and produce a prioritized list of required changes
• Apply AI-powered workflows and tools (e.g., Copilot, Sourcegraph Cody, linting assistants, testing frameworks) to refactor, optimize, and lint Python code, demonstrating improved correctness and maintainability as shown by passing tests and reduced lint warnings
• Demonstrate the use of platform-specific features (ChatGPT, Replit, Kaggle) to prototype, debug, and maintain code, including generating meaningful comments and annotations that explain intent and provenance
• Evaluate risks and trade-offs when blending manually written and AI-generated code by identifying sources of error, articulating mitigation strategies (testing, code review, provenance tracking), and documenting a maintenance plan