
| Course Code | : MCS514 |
| Course Type | : Area Elective |
| Couse Group | : Second Cycle (Master's Degree) |
| Education Language | : English |
| Work Placement | : N/A |
| Theory | : 3 |
| Prt. | : 0 |
| Credit | : 3 |
| Lab | : 0 |
| ECTS | : 6 |
This course aims to equip students with a foundational understanding of AI agents that perceive their environment reason, plan, and take actions to achieve goals. Students will progress from core concepts to hands-on code, exploring the design patterns, frameworks, and ethical considerations essential for building reliable and trustworthy AI agents using modern large language model (LLM) infrastructure.
This course introduces students to the theory and practice of building AI agents. It begins with an introduction to AI agents, their use cases, and the agentic paradigm. Students then explore major agentic frameworks and learn to identify and apply common agentic design patterns, including tool use, planning, multi-agent coordination, and metacognition. The course covers Agentic Retrieval-Augmented Generation (RAG) and techniques for building trustworthy, safe agents.
| 1. | Explain the concept of AI agents, their components, and common use cases. |
| 2. | Implement tool use, planning, and RAG patterns using Python and agentic frameworks. |
| 3. | Design and build a functional AI agent system. |
| 4. | Apply multi-agent design patterns to coordinate specialized agents for complex tasks. |
| 5. | Evaluate AI agent performance using systematic benchmarking and tracing methodologies. |
| 6. | Assess ethical, safety, and trust considerations in AI agent design. |
| 1. | Wang, L., Ma, C., et al. (2024). A Survey on Large Language Model based Autonomous Agents.Frontiers of Computer Science, 18(6), 186345.. |
| 2. | Xi, Z., et al. (2023). The Rise and Potential of Large Language Model Based Agents: A Survey. arXiv:2309.07864.Artificial Intelligence: A Modern Approach. |
| 3. | Weng, L. (2023). LLM-powered Autonomous Agents. Lilian Weng’s Blog. Retrieved from lilianweng.github.io/posts/2023-06-23-agent/. . |
| 4. | Yao, S., et al. (2023). ReAct: Synergizing Reasoning and Acting in Language Models. ICLR 2023. |
| 5. | Schick, T., et al. (2023). Toolformer: Language Models Can Teach Themselves to Use Tools. NeurIPS 2023. |
| 6. | Chase, H. (2023). LangChain Documentation. https://python.langchain.com |
| 7. | HuggingFace Datasets (tool-use, agent benchmarks) |
| 8. | AgentBench: A Comprehensive Benchmark for LLM-based Agents |
| 9. | ToolBench: Tool-use evaluation datasets (ToolLLM) |
| 10. | Python (autogen, langchain, langgraph) |
| 11. | GitHub |
| Type of Assessment | Count | Percent |
|---|---|---|
| Attending Lectures | 1 | %5 |
| Assignment | 1 | %10 |
| Midterm Examination | 1 | %15 |
| Final Examination | 1 | %70 |
| Activities | Count | Preparation | Time | Total Work Load (hours) |
|---|---|---|---|---|
| Lecture - Theory | 14 | 2 | 3 | 70 |
| Project | 2 | 20 | 0 | 40 |
| Presentation | 2 | 12 | 4 | 32 |
| Report | 2 | 6 | 0 | 12 |
| TOTAL WORKLOAD (hours) | 154 | |||
PÇ-1 | PÇ-2 | PÇ-3 | PÇ-4 | PÇ-5 | PÇ-6 | PÇ-7 | PÇ-8 | PÇ-9 | |
OÇ-1 | 3 | 3 | 3 | 3 | 5 | 3 | 5 | 3 | 3 |
OÇ-2 | 4 | 4 | 4 | 3 | 5 | 4 | 3 | 4 | 3 |
OÇ-3 | 3 | 3 | 5 | 5 | 5 | 4 | 3 | 5 | 5 |
OÇ-4 | 5 | 4 | 5 | 4 | 4 | 4 | 5 | 5 | 4 |
OÇ-5 | 3 | 3 | 3 | 3 | 4 | 3 | 4 | 3 | 3 |
OÇ-6 | 1 | 2 | 3 | 3 | 2 | 2 | 3 | 3 | 2 |