VVZ API is not affiliated with ETH Zurich. Data might be outdated or incorrect. Please view the official ETHZ Vorlesungsverzeichnis for binding information.

363-1209-00L 3 Credits MSC , NDS D-MTEC

Building Agentic AI Systems for Industry Transformation

Lecturers & Examiners: Dr. Markus Kreft, Dr. Kevin O'Sullivan, Dr. Robert Jakob
VVZ CR n/a

Last Updated: 2026-06-03 00:07:52

Abstract

AI is transforming industries through agentic systems. This course equips students with the engineering skills to design, build, and evaluate agentic systems, covering LLM fundamentals and context engineering through to multi-agent orchestration. Learning is delivered through lectures, guest speakers, and hands-on projects with direct relevance to real-world AI adoption.

Objective

This course introduces students to the core engineering and strategic building blocks of agentic AI systems, within the context of AI transformation in industry. After completing the learning modules below, students will be able to design, build, and evaluate agentic AI systems, while developing the practical and conceptual skills needed to identify and drive high-value AI adoption in real-world organizations. The learning modules will enable students to: • Design agentic AI systems by selecting appropriate architectures, from single-agent pipelines to multi-agent orchestration, based on real-world problem requirements. • Build core agentic components including context engineering pipelines, RAG systems, MCP integrations, and tool-using agents using standard frameworks and APIs. • Evaluate agent performance and reliability. • Distinguish between AI workflows and AI agents, and select the appropriate paradigms (e.g., fine-tuning, RAG, SLMs) for a given use case. • Identify high-value opportunities for agentic AI adoption within organizations and translate them into concrete technical proposals.

Content

This course is designed for Master's students with a technical background who are motivated to understand and build the agentic AI systems that are transforming industries and organizations. Students should have working knowledge of Python. No prior experience with large language models or AI frameworks is required. Agentic AI is rapidly becoming the dominant paradigm for applying AI in industry. Engineers who can build, evaluate, and deploy such systems are in growing demand. This course bridges the gap between academic AI knowledge and the practical engineering skills required to deliver agentic AI solutions in real organizational contexts. The course is organized into five thematic modules, delivered through weekly lectures, hands-on lab sessions, and guest speakers from industry: 1. Foundations: LLM basics, context engineering, prompt engineering, structured outputs, and the distinction between AI workflows and AI agents. 2. Context Engineering — Input Side: Retrieval-Augmented Generation (RAG) and the Model Context Protocol (MCP) as methods for grounding agents in external knowledge and data. 3. Context Engineering — Output Side: Structured output, tool use, function calling, planning, orchestration, and multi-agent system design. 4. Evaluation & Reliability: Agent evaluation, observability, safety, guardrails, and human-in-the-loop design. 5. AI Transformation in Industry: Use case discovery, organizational adoption, and AI governance. The course is assessed through a single semester project, completed in groups. Teams design, build, and deploy an end-to-end agentic AI system addressing a real-world industry use case, with final presentations held before a panel of academic and industry judges.

General Information

Language
English
Levels
MSC , NDS
Frequency
Yearly recurring

Examination

Type
ungraded semester performance
The primary deliverable of this course is a semester project in which groups of students identify, design, and build an agentic AI system addressing a real-world industry use case. Teams are expected to apply concepts introduced throughout the course in an integrated, end-to-end system, and present their work through a final presentation and live demo.

Course Components

Type Title Time & Place Hours
lecture with exercise Building Agentic AI Systems for Industry Transformation No time listed 2 h semesterly

Offered In