Building AI Agents: A Hands-on Workshop

Description:

This full-day hands-on workshop provides a deep dive into Agentic AI, their architecture, and real-world applications. Participants will build AI Agents with advanced capabilities, including multi-agent collaboration, memory, and task automation using open-source tools.

Duration: 1 Day 

Course Code: BDT402

Learning Objectives:

After this course, you will be able to:

• Implement complex AI Agent systems with memory management and tool integration
• Develop multi-agent systems that can collaborate and communicate effectively
• Design and integrate custom tools and capabilities into AI Agent frameworks
• Build, test, and deploy production-ready AI Agent applications with appropriate error handling and security considerations

• Proficiency in Python
• Familiarity with APIs and LLMs (e.g., OpenAI, Hugging Face)
• Basic understanding of AI concepts

• Developers
• AI engineers
• Students with Python and LLM experience

Course Outline:

• Overview of autonomous agents
• Agent design patterns and architectures
• Architectures: Reflex, Deliberative, Hybrid
• Planning systems and decision-making
• Advanced memory architectures
• State management and persistence
• LangChain, CrewAI, AutoGen frameworks

• Setting up a development environment
• Installing necessary libraries (LangChain, AutoGen, ChromaDB, etc.)
• Connecting AI Agents to APIs and knowledge bases
• Implementing agent framework from scratch
• Creating custom tools and capabilities
• Advanced prompt engineering techniques

• Vector database implementation
• Long-term and working memory
• Context window management
• Hierarchical memory structures

• Creating custom tools
• API integration patterns
• Function calling and tool use
• Error handling and recovery

• Agent communication protocols
• Implementing agent cooperation
• Task distribution and management
• Conflict resolution

• Building a complete agent system
• Testing and evaluation
• Performance optimization
• Deployment considerations

• Security considerations
• Scaling agent systems
• Latest research and trends
• Resources for continued learning

• Building a task management agent
• Implementing a research assistant
• Creating a multi-agent system