Agents and Agentic Workflows: A Hands-On Introduction

---

Oct 3 through Oct 24, 2026. Four Saturdays 1:30-4:30pm (10/3, 10/10, 10/17, 10/24).

The IEEE North Jersey Section Communications Society Chapter is offering a course entitled
"Agents and Agentic Workflows:  A Hands-On Introduction". 

This four-week introductory course in Agents designed for a technical audience 
with no prior specialization in AI or machine learning. It provides a practical, 
hands-on approach to understanding how Agent Workflows operate and how they can be applied across tasks.

Agents are autonomous software systems that use a large language model in order to perceive their environment, make decisions, and take actions toward a specific goal.  Unlike standard chatbots that just answer questions, agents can reason, plan and use external tools independently.  This course is a four-week introduction to building agent-based applications designed for a general technical audience with no prior specialization in AI or machine learning.

Each week combines a lecture with interactive demos and assignments to reinforce learning through real-world use cases. It provides a practical, hands-on approach to implementing agentic workflows using the LangChain platform.

The IEEE North Jersey Section's Communications Society Chapter can arrange for providing IEEE CEUs - Continuing Education Units (for a $5 charge) upon completion of the course.  Course prices: $75 for Undergrad/Grad/Life/ComSoc members, $100 for IEEE members, $150 for non-IEEE members.  If paying by check, make payee out to "IEEE North Jersey Section".

---

---

---

---

Agenda: The primary objective of this course is to provide students with an understanding of Agents, tools and techniques used, the wide variety of applications, and an Agentic future.
The material covered includes an introduction to the concepts and how to build applications using these concepts.  On the completion of the course, students will learn:

Week 1: Generative AI Agents and Autonomous Workflows
Theme: Overview of AI agents, their architecture, and how they orchestrate autonomous tasks using LLSs.
Outline:
-Whare are AI agents ?
-Review of important LLM concepts
-LLMs as the brains of agents
-Overview of how agents use tools, memory, and planning
-Introduction to LangChain

Week 2: Tools
Theme: Empowering agents with external tools
Outline:
-Tools are the callable functions that allow an LLM to do something in the "real world"
-Customize the capabilities of the LLM with external tools
-Creating tools in LangChain:  name, description, arguments
-The agent-tool loop and execution
-Built-in tools (harness) vs. custom tool creation
-Subagents as tools
-Error handling

Week 3: Model Context Protocol
Theme: Connecting agents to proprietary data sources and systems.
Outline:
-MCP is a standardized way for agents to securely access proprietary databases and APIs
-Client-server architectures:  agent, client, server
-Moving beyond tools:  resources (data streams) and prompts (templates generated by the server)
-Security issues:  client-server architecture allows controlled access to sensitive data

Week 4: Skills
Theme: Step-by-step procedures for domain specific tasks
Outline:
-Skills encode workflow, judgment, and step-by-step reasoning
-SKILL.md files
-Progressive disclosure:  availability of skills based on context or user needs
-Conditional loading: load only the right procedure at the right time to avoid context bloat
-Reusability:  the same skill format can be used across multiple projects

Technical Requirements :

Any coding demos in this course will use the Python programming language and will be distributed in the form of Colab notebooks.  Since Colab is a cloud-based platform, participants will not need to install any software locally on their own devices.

The main platform used during this course will be LangChain, an open-source framework for building agents.  At the heart of any agentic workflow is a Large Language Model.  This course will specifically demonstrate how to incorporate Gemini into LangChain agents.  However, LangChain treats the LLM as a swappable component of the workflow.  Participants with access to other LLMs, such as GPM or Claude, are welcome to incorporate those models instead.

Basic programming skills and some familiarity with the Python language are assumed.
Students are expected to be able to bring a laptop in order to use Google Colab Notebooks.

The course is intended to be subdivided into 4 sessions, each three hours long for a total of 12 course hours. Each lecture is further subdivided into lecture, guided and independent project based exercises to build experience with hands-on techniques.  CEUs will be made available.

This course will be held at FDU - Teaneck, NJ campus.  Checks should NOT be mailed to this address. Can physically bring (preferred) checks in person on the first day or use online payments at registration. Email the organizer for any questions about course, registration, or other issues.

---

Free Parking available in the main North Lot (large parking lot with lots of spaces, only use the public spaces provided) right off Route 4 entrance.

---