Databricks GenAI Advanced Lab

A practical 2-day hands-on workshop for building enterprise-level AI agents using Databricks platform capabilities.

Key Learning Outcomes

• Build Agent Systems: Create single-agent and multi-agent architectures with vector search, function calling, and tool integration
• Master LLMOps Pipeline: Implement the complete lifecycle from development to production including evaluation, monitoring, deployment, and Databricks Apps for front-end applications
• Compare Agent Approaches: Hands-on experience with different architectures (single-agent vs multi-agent vs MCP), prompt optimization, and quick deployment approaches like Agent Bricks and Genie spaces

Prerequisites

• Databricks workspace access and compute (Serverless or Classic)
• Basic Python knowledge

Lab Structure

Day 1: Foundational Agent Building

Setup & Data Exploration

• 00_setup/ - Environment configuration and data preparation
• 01_explore_data/ - Data analysis and preprocessing techniques

Vector Search & Tools

• 02_create_vector_search_index/ - Building semantic search capabilities
• 03_create_tools/ - Creating custom function tools for agents

Agent Development

• 04_create_agent_with_vsi_and_tools/ - Core agent with vector search integration
• 05_eval_agent_and_deploy/ - Agent evaluation and deployment strategies

Production Readiness

• 06_setup_sme_review/ - Subject matter expert review workflows
• 07_monitor_agent_in_production/ - Production monitoring and observability

User Interfaces

• 08_create_chatbot_app/ - Web-based chatbot application

Day 2: New & Advanced Features

Multi-Agent Development

• 09_create_genie_space (UI)/ - Databricks Genie space integration
• 10_create_multi_agent_with_tools_and_genie (ChatAgent)/ - Multi-agent systems

Advanced Architectures

• 11_create_agent_with_mcp/ - Model Control Protocol integration
• 12_create_agent_bricks (UI)/ - Agent Bricks UI components
• 13_prompt_optimization/ - Advanced prompt engineering techniques

Data

The data/ directory contains sample ecommerce datasets:

• CSV files: customer services, product documentation, policies, inventories
• PDF files: sample product manuals for hands-on document processing

Getting Started

1. Run 00_setup/00_setup.py to initialize your environment
2. Follow modules sequentially for best learning experience
3. Each module builds upon previous concepts