| Lehrinhalt |
"Generative AI in Software Development" - Building software with AI
agents responsibly and rigorously
Examination:
Graded projects & written exams
Course description:
Block 1 – Foundations: Java/with Spring, Spring Boot, and Spring Data
JPA fundamentals before covering how LLMs generate text, prompt
engineering, RAG and vector databases, agentic workflows, and MCP. Labs:
Java with Spring AI and Embabel.
Block 2 – Building Good Code, Catching Bad Code: Learn and apply SOLID
principles and common design patterns as concrete criteria for judging
AI-generated code, cope with AI code smells & common security
issues/guardrails, students learn to judge (AI-generated) code against
these principles.
Block 3 – From Reviewer to Delegator: Introduction to an agentic coding
tool: Claude Code Pro (paid) or Antigravity CLI (free, requires a Google
account) - project context files, skills, hooks, subagents,
permission/autonomy modes; SDD (Spec-driven development) with Spec-Kit
Block 4 – Final Project & Examination: an individually built,
spec-driven application, which will be graded on planning artifacts,
code quality, and process evidence, and a closed-book written exam with
project-specific and theory questions.
*Important note: From Block 3 onward, students need either a paid Claude
Code Pro subscription or a Google account for the free Antigravity CLI
track.* |
