Building AI Agents using Generative AI
Duration – 2 Days
Objectives
- Understand AI agents core concepts
- Explore LLM’s and Prompt Engineering
- Set up tools and frameworks
- Build memory-enabled intelligent agents
- Develop tools using AI applications
Tools & Platforms
- LangChain Python SDK
- AgentLite GitHub Toolkit
- LlamaIndex Document Indexing
- VS Code with Python
- OpenAI or Hugging Face API
- Build Q&A Document Agent
Pre-requisites
- Basic understanding of Python programming.
- Familiarity with API usage
- Understanding of Large Language Models
Outcomes
- Understand AI-Agent Architecture
- Build LLM powered simple agents
- Integrate tools into AI agents
- Build memory driven work flows
- Deploy a simple agent
Day 2: Building and extending AI Agents
Agent workflows and memory
- What is memory in AI agents?
- Types of memory: Buffer, Summary, VectorStore
- Add memory to your agent using LangChain
Agent Tools and Action
- What is ReAct and why it is important in Agentic AI?
- Utilizing external tools such as search engines, calculators, and code interpreters
- Connecting and working with APIs like Wikipedia, weather services, and web search platforms
- Develop an agent that uses tools to respond to real-world questions
Project ideas
- Student Assistant (Q&A from syllabus)
- Resume Reviewer Bot
- Code Helper Bot using Python docs
