AI Agent Workshop 🚀

Welcome to the AI Agent Workshop. In this hands-on course, you will learn how to architect, build, and deploy production-grade AI Agents using the Google Agent Development Kit (ADK) and Gemini models.

Project Overview

You will build a Autonomous Financial Research Team designed for enterprise use. This system evolves from a simple script into a sophisticated, multi-agent organization capable of complex financial reasoning, deep research, and executive reporting.

The Agent Evolution

The workshop follows a progressive "Crawl, Walk, Run" methodology, where you build layers of capability module by module:

Phase 1: The Specialist (Module 1 & 2)

You start by building individual experts:

• The Quant: connected to live market data tools to fetch stock prices and perform financial math.
• The Analyst: grounded in internal knowledge (Investment Policies, Strategy Papers) using Retrieval-Augmented Generation (RAG).

Phase 2: The Manager (Module 3)

You introduce hierarchy:

• The Manager: A routing agent that understands user intent and delegates tasks to the appropriate specialist, synthesizing their outputs into a coherent answer.

Phase 3: The Organization (Module 4)

You implement complex orchestration:

• The Autonomous Team: A full research department that can take a high-level goal (e.g., "Analyze our AI strategy"), break it down, perform iterative research loops with self-correction and compliance checks, and produce a polished executive memo.

Workshop Structure

Module	Topic	Key Concepts
Module 1	The Zero-to-One Agent	Tools, Function Calling, Basic Reasoning
Module 2	RAG-Powered Agents	Vector Ops, Embedding, Grounding, MCP
Module 3	Workflow Agents	Hierarchical Delegation, Routing, State
Module 4	Complex Multi-Agent Systems	Sequential & Loop Flows, Custom Logic, Self-Correction

Getting Started

1. Prerequisites

• Python 3.10+
• A Google AI Studio API Key

2. Environment Setup

1. Clone the repository:

   git clone https://github.com/heiko-hotz/jpmorgan-ai-agent-workshop.git
   cd jpmorgan-ai-agent-workshop

2. Create a virtual environment:

   python3 -m venv .venv
   source .venv/bin/activate

3. Install dependencies:

   pip install google-adk google-genai python-dotenv chromadb

4. Configure API Key: Create a .env.local file in the root directory:

   GOOGLE_API_KEY=your_api_key_here

3. Running the Workshop

Each module contains a Jupyter Notebook (.ipynb) for learning and a Python folder (e.g., financial_agent_app/) for the production-ready code.

Start with Module 1:

jupyter notebook module_01/01_fast_track_agent.ipynb

Technologies Used

• Google ADK: Framework for agent construction, orchestration, and evaluation.
• Gemini 2.5 Flash: High-performance, cost-effective LLM for reasoning.
• ChromaDB: Open-source embedding database for RAG.
• Model Context Protocol (MCP): Standard interface for connecting AI models to data.