Building AI Agents
summary
Walks through building a working agent that moves from VS Code prototype to Azure deployment using AI Toolkit and GitHub Copilot, with live model recommendations (GPT‑4.1 Mini, GPT‑5 Nano, Claude Haiku/Opus/Sonnet), an MCP server (TypeScript or Python template) providing mock weather/activity tools, a model playground to compare responses, auto‑generated agent code, traced runs visible in the tracing viewer showing tool invocations, and an automated evaluation workflow; the workshop completes in about one hour.
AI-written summary · about
chapters (14)
- 0:00Welcome & Workshop Overview: AI Toolkit + GitHub Copilot
- 0:20Why Combine AI Toolkit with Copilot? Key Benefits
- 1:01Setting Up Your Environment: Foundry Project & VS Code Extensions
- 2:32Connecting Foundry to VS Code: Default Project Setup
- 4:41GitHub Copilot Agent Mode: Recommended Models & Setup
- 5:36Demo: Getting Model Recommendations with AI Toolkit Tools
- 9:19Testing Models in AI Toolkit Playground: Compare & Choose
- 12:34Building Your Agent: Using Agent Builder in VS Code
- 14:08Creating MCP Servers: Templates & Workflow in AI Toolkit
- 16:05Adding Tools & Improving System Prompts for Your Agent
- 19:06Migrating to Code: Generating Agent Code from Agent Builder
- 21:19Enabling Tracing with GitHub Copilot & Viewing Traces
- 23:02Adding Evaluations: Automated Workflow & Reports
- 24:17Workshop Resources & Wrap-Up: AI Toolkit Samples Repo
related talks
23mFine Tune Your AI Toolbox with GitHub Copilot, Azure AI Foundry...
36mBuilding AI Agents in .NET with Azure AI Foundry
39mEvolving Intelligence: Designing Autonomous AI Agents
20mAgentic Excellence: Mastering AI Agent Evals w/ Azure AI Evaluation SDK
58mAdvanced Autonomous AI Agents: Memory, Collaboration and Evolution in Action
01h 23mBuild your own AI Agent
01h 30mCreating Intuitive AI Agents for PowerShell..
56mBuilding AI Agents and Multi-Agent Systems with Azure AI