SI 405: Applied Generative AI

Course Description

This course offers a practical, hands-on exploration of generative artificial intelligence, specifically focusing on large language models (LLMs) and their real-world applications. Students will gain experience building applications using state-of-the-art LLMs via cloud APIs (such as OpenAI GPT, Anthropic Claude, or DeepSeek) and running local models using tools like Ollama or Hugging Face Transformers.

The curriculum prioritizes applied techniques—including retrieval-augmented generation (RAG), agentic AI, and multi-agent social simulations—over deep mathematical foundations or theory. All development is conducted in Python.

Learning Objectives

By the end of the semester, students will be able to:

• Implement RAG: Integrate external knowledge bases into LLM-based systems.
• Build Tool-Using AI Agents: Develop agentic systems where an LLM reasons about tasks and uses external tools or APIs (e.g., calculators, search) to solve problems.
• Simulate Multi-Agent Interactions: Model social scenarios or collaborative tasks using multiple interacting LLM-driven agents.
• Apply Modern Frameworks: Use libraries like LangChain to streamline pipelines and memory management.
• Evaluate Outputs: Use metrics and human judgment to address hallucinations and response errors.
• Address Ethics & Security: Understand model limitations (bias, privacy) and best practices for responsible deployment.

Required Skills

To succeed in this course, students should possess:

• Moderate Python Proficiency: Ability to write scripts, handle data structures, and use libraries.
• API Familiarity: Experience with RESTful APIs and JSON formats.
• Basic ML/AI Knowledge: Understanding concepts like training vs. inference.
• Prior LLM Exposure: Basic experience with prompt engineering and pre-trained models.
• Debugging Basics: Fundamental understanding of code debugging and model deployment.

Course Schedule (Winter 2026)

Date	Lecture	Topic	Due
1/7/2026	1	Course Overview & Generative AI Landscape
1/12/2026	2	Tools and Platforms for Generative AI Development
1/14/2026	3	Introduction to LangChain
1/21/2026	4	Building LLM Chains and Memory with LangChain
1/26/2026	5	Text Embeddings and Vector Stores for LLMs
1/28/2026	6	Incorporating External Knowledge via RAG
2/2/2026	7	Building Robust QA Systems with RAG
2/4/2026	8	Evaluating RAG systems and Output Quality	Homework 1
2/9/2026	9	Advanced RAG: Hybrid Search, Reranking, and GraphRAG
2/11/2026	10	Introduction to LLM Agents and Tool Use
2/16/2026	11	Agentic AI Concepts and Planning
2/18/2026	12	Frameworks for Building LLM Agents	Homework 2
2/23/2026	13	Integrating Web Search, Services, and APIs
2/25/2026	14	Midterm
3/9/2026	15	Integrating Calculation & Code Execution
3/11/2026	16	Coordinating Multiple Tools & Planning Strategies	Project Proposal
3/16/2026	17	Memory and Context Management for AI Agents
3/18/2026	18	Multimodal Agents: Vision, Voice, and Documents
3/23/2026	19	Advanced and Autonomous Agent Systems
3/25/2026	20	Evaluating and Debugging AI Agents
3/30/2026	21	Ethical and Societal Implications	Homework 3
4/1/2026	22	Security and Safety in Agentic Systems
4/6/2026	23	Aligning Models during Inference and Training
4/8/2026	24	Social Simulation using LLM Agents
4/13/2026	25	Designing Personified Agents: Identity & Memory
4/15/2026	26	Emergent Social Dynamics & Practical Deployment	Homework 4
4/20/2026		Project Presentations

Grading and Coursework

Grades are determined by a combination of individual assignments and a final project.

Grade Breakdown

• Weekly Mini-Assignments: 20%
• Large Assignments: 30%
• Course Project: 30%
• Midterm Exam: 10%
• Final Exam: 10%

Grading Scale

The course uses a traditional scale (A+ to F). There is no curve. No extra credit or make-up work is permitted, though instructors may adjust a borderline grade for an exceptional project.

Academic Integrity

Unless otherwise specified in an assignment all submitted work must be your own, original work. Any excerpts, statements, or phrases from the work of others must be clearly identified as a quotation, and a proper citation provided. Any violation of the School's policy on Academic and Professional Integrity (stated in program-specific student handbooks) will result in serious penalties, which might range from failing an assignment, to failing a course, to being expelled from the program. Violations of academic and professional integrity will be reported to UMSI Office of Academic and Student Affairs. The faculty instructor determines consequences impacting assignment or course grades; the School may impose additional sanctions.

• Original Work: Unless otherwise specified, all work must be original and cited properly.
• Generative AI Policy: You are encouraged to use tools like ChatGPT, Cursor, and Co-Pilot for assignments and projects. However, in-class coursework must be done by humans to ensure proper skill development.
• Collaboration: Sharing assignment code publicly on GitHub is a violation of the Honor Code.

Accommodations for Students with Disabilities

If you think you need an accommodation for a disability, please let me know at your earliest convenience. Some aspects of this course, the as­signments, the in-class activities, and the way we teach may be modified to facilitate your participation and progress. As soon as you make me aware of your needs, we can work with the Office of Services for Students with Disabilities (SSD) to help us determine appropriate accommoda­tions. SSD (734-763-3000; ssd.umich.edu) recommends students request disability-related academic accommodations via the Accommodate system, a core electronic case management system that will assist students, faculty, instructors, and staff in requesting, approving, and implementing disability-related accommodations. I will treat any information that you provide in as confidential a manner as possible.

Student Wellbeing

Students may experience stressors that can impact both their academic experience and their personal well-being. These may include academic pressure and challenges associated with relationships, mental health, alcohol or other substances, identities, finances, food insecurity, or other external stressors.

If you are experiencing concerns, seeking help is a courageous thing to do for yourself and those who care about you. If the source of your stressors is academic, please contact UMSI's academic success team via umsi.academicsuccess@umich.edu or me so that we can find solutions together.

For a full list of resources, please see the UMSI Academic Success page or the PDF version of the syllabus