Course Syllabus

Agentic Artificial Intelligence

Course Code: AAI-PhD Semester: Spring 2025 Credits: 6 ECTS Format: Weekly lectures + hands-on labs


Course Description

This doctoral-level course provides a comprehensive introduction to agentic artificial intelligence - autonomous systems that leverage large language models to reason, plan, use tools, and collaborate. Students will learn both theoretical foundations and practical implementation skills.

Learning Outcomes

Upon completion, students will be able to:

  1. Design agent architectures for complex reasoning tasks
  2. Implement tool-using agents with multiple LLM providers
  3. Build multi-agent systems with effective coordination
  4. Evaluate agent performance using standard benchmarks
  5. Apply safety measures to prevent hallucinations
  6. Create domain-specific agents for real-world applications

Prerequisites

  • Strong Python programming skills
  • Familiarity with machine learning concepts
  • Basic understanding of transformer architectures
  • Ability to read and understand research papers

Required Materials

  • Access to OpenAI or Anthropic API (student credits available)
  • Python 3.11+ development environment
  • GitHub account for coursework submission

Schedule

Week Topic Deliverables
1 Introduction to Agentic AI Lab 1: First Agent
2 LLM Foundations for Agents Lab 2: Prompting Strategies
3 Tool Use and Function Calling Lab 3: MCP Implementation
4 Planning and Reasoning Lab 4: Reflexion Agent
5 Multi-Agent Architectures Lab 5: Agent Team
6 Agent Frameworks Lab 6: Framework Comparison
7 Advanced RAG Lab 7: Self-RAG
8 GraphRAG and Knowledge Lab 8: Knowledge Graph
9 Hallucination Prevention Lab 9: Fact Checker
10 Agent Evaluation Lab 10: Benchmark Suite
11 Domain Applications Lab 11: Domain Agent
12 Research Frontiers Final Project Presentation

Assessment

Component Weight Description
Weekly Labs 40% Hands-on implementation exercises
Paper Presentations 20% Critical analysis of assigned papers
Final Project 30% Novel agent system development
Participation 10% Class discussions and peer review

Grading Scale

Grade Percentage
A 90-100%
B 80-89%
C 70-79%
D 60-69%
F Below 60%

Policies

Late Submissions

  • Labs accepted up to 48 hours late with 20% penalty
  • No late submissions for final project

Academic Integrity

  • All code must be original or properly attributed
  • Use of LLM assistants must be disclosed
  • Collaboration encouraged but individual submissions required

Accommodations

  • Students needing accommodations should contact the instructor within the first week

Office Hours

Instructor: Prof. Dr. Joerg Osterrieder Email: joerg.osterrieder@fhgr.ch Office Hours: By appointment via email

Resources


This syllabus is subject to change. Students will be notified of any modifications.


Back to top