Designing AI Agents with LLMs and VLMs
A project-based laboratory course on agent loops, tools and MCP, retrieval, multimodal planning, safety, and evaluation.
Course Format
This course consists of five six-hour laboratory meetings. Each student develops an individual agent project during Meetings 1–4 and presents its final version in Meeting 5. All course materials and project work are in English.
Start with Project Definition and Agent Loop and create your repository from the course template.
Assignments
- Project Definition and Agent Loop
- Tools, Function Calling, and MCP
- State, Memory, RAG, and Execution Graphs
- Vision, Planning, Safety, and Evaluation
- Individual Project Presentations
Assessment
The final integrated project and its presentation account for 100% of the course grade.