Course Syllabus

Agentic Large Language Models

Course Code: TBD
Credits: TBD
Institution: LIACS, Leiden University

Course Description

This course provides a comprehensive introduction to building autonomous agents powered by Large Language Models (LLMs). Students will learn the theoretical foundations, practical techniques, and ethical considerations for developing intelligent systems that can reason, plan, and act independently.

Prerequisites

  • Basic programming skills (Python recommended)
  • Fundamental understanding of machine learning concepts
  • Familiarity with neural networks (helpful but not required)

Learning Outcomes

Upon successful completion of this course, students will be able to:

  1. Explain the architecture and capabilities of state-of-the-art Large Language Models
  2. Design and implement autonomous AI agents using LLM frameworks
  3. Apply advanced prompt engineering techniques for complex reasoning tasks
  4. Integrate LLMs with external tools, APIs, and knowledge bases
  5. Build and coordinate multi-agent systems for collaborative problem-solving
  6. Evaluate agent performance and ensure responsible AI deployment
  7. Critically assess the limitations and risks of agentic AI systems

Course Topics

Module 1: Foundations of Large Language Models

  • Transformer architecture and attention mechanisms
  • Pre-training and fine-tuning strategies
  • Capabilities and limitations of current LLMs
  • In-context learning and few-shot prompting

Module 2: Prompt Engineering

  • Prompt design principles and best practices
  • Chain-of-thought reasoning
  • Self-consistency and voting mechanisms
  • Prompt optimization techniques

Module 3: Building AI Agents

  • Agent architectures (ReAct, Plan-and-Execute, Reflexion)
  • Memory systems and context management
  • Tool use and function calling
  • Action planning and execution

Module 4: Advanced Agent Capabilities

  • Multi-agent systems and collaboration
  • Self-reflection and iterative improvement
  • Long-term planning and goal decomposition
  • Handling uncertainty and errors

Module 5: Integration and Deployment

  • Connecting to external APIs and databases
  • RAG (Retrieval-Augmented Generation) systems
  • Agent orchestration frameworks
  • Production deployment considerations

Module 6: Ethics, Safety, and Evaluation

  • Responsible AI principles
  • Bias detection and mitigation
  • Safety constraints and guardrails
  • Evaluation metrics for agent systems

Assessment

  • Assignments: 40% - Weekly programming exercises and problem sets
  • Midterm Project: 20% - Design and implement a simple agent system
  • Final Project: 30% - Build a comprehensive agentic application
  • Participation: 10% - Active engagement in discussions and labs

Required Materials

  • Lecture slides and notes (provided online)
  • Selected research papers and articles
  • Access to LLM APIs (OpenAI, Anthropic, or open-source alternatives)
  • Python programming environment with relevant libraries
  • Research papers on LLMs and agent systems (provided throughout the course)
  • “Prompt Engineering Guide” (online resource)
  • Documentation for popular LLM frameworks (LangChain, AutoGen, etc.)

Schedule

See the Schedule page for detailed week-by-week topics and assignment deadlines.

Academic Integrity

Students are expected to follow Leiden University’s academic integrity policies. While collaboration on concepts is encouraged, all submitted work must be original. Proper attribution is required for any external resources used.

Accessibility

If you require any accommodations to participate fully in this course, please contact the instructor or university disability services as soon as possible.

Course Updates

This syllabus may be updated during the course. Students will be notified of any significant changes.

Last updated: February 2026