Quests
Complete at least 2 quests (including the guided project) to be eligible for the hack. Each quest teaches a core AI development pattern with JavaScript/TypeScript.
💻 Quest 1: Local AI Development with Foundry Local
Run AI models on-device for privacy, speed, and offline access using Microsoft Foundry Local.
You'll learn to:
- Install and manage local AI models with Foundry Local CLI
- Use the Foundry Local SDK with LangChain.js
- Build an AI Insight Mapper that extracts structured data from unstructured text
- Power GitHub Copilot with local models for private coding assistance
Award Tie-in: Offline-Ready AI Award
This quest directly maps to the Offline-Ready AI Award category. Build an app that processes sensitive data entirely on-device and highlights privacy + latency optimizations.
☁️ Quest 2: End-to-End Model Development on Microsoft Foundry
Build AI solutions end-to-end in the cloud with Microsoft Foundry models and workflows.
You'll learn to:
- Explore model selection workflows using catalog, leaderboards, and playgrounds
- Understand customization options like fine-tuning
- Evaluate model quality, safety, and trustworthiness
- Apply cloud-scale AI development patterns for real-world scenarios
🔍 Quest 3: Run a Serverless RAG Support System
Build a retrieval-augmented generation pipeline with LangChain.js and local models.
You'll learn to:
- Set up a serverless RAG system with document ingestion
- Implement vector storage and semantic retrieval
- Generate grounded responses with citations
- Handle streaming and chat history
🤖 Quest 4: Build Agents with AI Toolkit
Create AI agents with the AI Toolkit for VS Code — from model discovery to evaluation.
You'll learn to:
- Explore and compare models in the Model Catalog
- Build agents with Agent Builder and dynamic variables
- Integrate MCP servers for tool use
- Evaluate agent performance with built-in and custom metrics
- Export production-ready code
Award Tie-in: Agentic System Architecture Award
This quest teaches the patterns you need for the Agentic System Architecture Award — innovative agent design with MCP tool integration.
🍔 Quest 5: Run an E2E Burger Ordering Agent System
Design and run a full agentic system with MCP server integration using LangChain.js.
You'll learn to:
- Understand AI agent architecture patterns
- Design multi-service agent workflows
- Build MCP tools for external system integration
- Create agent APIs and backend services
- Plan for production scaling with Azure
Award Tie-in: Agentic System Architecture Award
This quest demonstrates foundational patterns for useful, reliable, and secure agent systems.