Foundations of AI Agents

A six-week graduate elective on what AI agents are, how to build them, and what it takes to put them in front of real users. Lab-driven and designed for students who learn by building.

Designed by Srikanth Jagabathula, Ilan Lobel, Raveesh Mayya, and Omar Mouchtaki · NYU Stern, 2026

See the curriculum → Student projects

About the course

A deep dive into the
craft of building agents.

Foundations of AI Agents is a graduate elective for MBA students at NYU Stern. Over six three-hour sessions, students move from the fundamentals of how agentic systems work (tool calling, system prompts, model trade-offs, managing stochasticity) into the design and orchestration of multi-agent systems, retrieval-augmented enterprise agents, and rigorous evaluation.

The course is built around the conviction that agents are best learned by building them. Every session pairs concept with lab: students leave each class with an agent they wrote themselves. The term culminates in Demo Day, where teams present agents they designed for real business problems. Offered starting in Spring 2026.

Curriculum

Six sessions, lecture and lab.

Each Tuesday for six weeks, three hours per session, with concept and build time interleaved. Days 1-4 are technical; Day 5 brings in practitioners; Day 6 is the students' show.

01Day

Building an AI agent

What an AI agent is, how LLMs work, tool calling and JSON, system prompts, model trade-offs, stochasticity management, the value of orchestration.

Lecture + lab
02Day

Agent architecture

A design template for agents, system-prompt design, hard-coded vs. LLM-determined behavior, multi-agent architecture.

Lecture + lab
03Day

The power of agentic AI

Agentic coding, building a frontend for an AI agent, security considerations, "MBA as a vibecoder": how non-engineers can ship real software with the right scaffolding.

Lecture + lab
04Day

Enterprise agents and evaluation

Context windows, embeddings, retrieval-augmented generation, and the harder question: how do you actually measure whether an agent is any good?

Lecture + lab
05Day

Deploying AI agents

An AI case on the challenges of deploying AI agents within organization, followed by a guest panel of practitioners deploying AI agents in production: what does the road from prototype to deployment actually looks like?

Case + guest panel
06Day

Demo Day

Student teams present agents they designed and built for real business problems. Peer evaluation, instructor review, and a showcase of the term's work.

Showcase

Team

Designed and taught at NYU Stern.

The course was jointly designed by four faculty in NYU Stern's Technology, Operations, and Statistics department, drawing on their work at the intersection of operations and AI.

Srikanth Jagabathula

Srikanth Jagabathula

Professor of Technology, Operations and Statistics, NYU Stern
Ilan Lobel

Ilan Lobel

Professor of Technology, Operations and Statistics, NYU Stern
Raveesh Mayya

Raveesh Mayya

Assistant Professor of Technology, Operations and Statistics, NYU Stern
Omar Mouchtaki

Omar Mouchtaki

Assistant Professor of Technology, Operations and Statistics, NYU Stern