Lab 4 – Build Meeting Assistant - Your First AI Agent with OpenAI SDK | Using Tools & Tracing 🤖

Description

Welcome to Lab 4 of the Agentic AI Series! 🎓 In this session, we take a huge leap — from simple LLM calls to creating your first real Agent using the OpenAI SDK. 🚀 We’ll build an AI Meeting Assistant that can: ✅ Summarize meeting transcripts ✅ Extract key action items with owners ✅ Send notifications via SendGrid and Pushover ✅ Automatically decide which tools to use through the Agentic workflow 🧠 What You’ll Learn: What is an Agent and how it differs from a regular LLM Components of an Agent: Tools, Resources, LLM, Memory, and Tracing How to define your own function-based tools using decorators How to integrate email and push notification APIs How to build a simple Gradio UI for your agent How to visualize execution flow using OpenAI Traces Resources: Github: https://github.com/TEJAPS/agentic-dpoint/tree/main ⏱️ Timestamps 00:00 Intro 02:00 What is an Agent 02:20 What is a Tool 03:00 Resources 06:00 Agent Workflow 07:00 OpenAI Agent SDK 08:00 SendGrid and Pushover Setup 12:47 Function Tools 14:22 Gradio Chatbot UI 15:18 LLM vs Agentic Tool 18:10 Function as Tool for Agent 23:30 Tracing 🧩 Tech Used Python 🐍 OpenAI SDK Gradio SendGrid API Pushover Notifications Agent SDK OpenAI Traces 💬 Join the Journey We’re building from Prompts → Agents → Agentic Systems step by step. Follow the series to learn how to design intelligent, tool-using AI systems that think and act independently. 📺 Previous Labs: Lab 1: OpenAI SDK Basics Lab 2: Multi-Model Comparison (Ollama vs GPT) Lab 3: AI Resume Reviewer