{"id":51330304,"url":"https://github.com/olivomarco/ai-hackathon","last_synced_at":"2026-07-01T22:04:32.351Z","repository":{"id":361016708,"uuid":"1252676683","full_name":"olivomarco/ai-hackathon","owner":"olivomarco","description":"Microsoft Foundry hackathons","archived":false,"fork":false,"pushed_at":"2026-05-28T20:30:07.000Z","size":803,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-05-28T22:12:55.787Z","etag":null,"topics":["ai","hackathon","microsoft-foundry"],"latest_commit_sha":null,"homepage":"https://olivomarco.github.io/ai-hackathon/","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/olivomarco.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-05-28T18:59:47.000Z","updated_at":"2026-05-28T20:30:38.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/olivomarco/ai-hackathon","commit_stats":null,"previous_names":["olivomarco/ai-hackathon"],"tags_count":null,"template":false,"template_full_name":null,"purl":"pkg:github/olivomarco/ai-hackathon","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/olivomarco%2Fai-hackathon","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/olivomarco%2Fai-hackathon/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/olivomarco%2Fai-hackathon/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/olivomarco%2Fai-hackathon/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/olivomarco","download_url":"https://codeload.github.com/olivomarco/ai-hackathon/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/olivomarco%2Fai-hackathon/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":35024386,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-26T15:22:16.424Z","status":"online","status_checked_at":"2026-07-01T02:00:05.325Z","response_time":130,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["ai","hackathon","microsoft-foundry"],"created_at":"2026-07-01T22:04:31.871Z","updated_at":"2026-07-01T22:04:32.346Z","avatar_url":"https://github.com/olivomarco.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Build Intelligent Apps with Microsoft Foundry\n\n[![Deploy GitHub Pages](https://github.com/olivomarco/ai-hackathon/actions/workflows/deploy-pages.yml/badge.svg)](https://github.com/olivomarco/ai-hackathon/actions/workflows/deploy-pages.yml)\n[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)\n[![Open in GitHub Codespaces](https://github.com/codespaces/badge.svg)](https://codespaces.new/olivomarco/ai-hackathon)\n\n*A What The Hack hackathon: from prompt to production*\n\n---\n\n## What is What The Hack?\n\n**What The Hack** (WTH) is Microsoft's hackathon-in-a-box format designed to teach cloud and AI technologies through hands-on challenge-based learning. This repository brings WTH to Microsoft Foundry, a unified platform for building, evaluating, and deploying intelligent applications.\n\nThis hackathon, you and your team build **one evolving artifact** — the **Northfield University IQ Assistant** — across three tiers. First you complete a guided **Foundations** challenge (four ordered steps) to stand up a deployed, grounded agent; then you pick **Advanced** challenges in any order to make it act, prove it safe, observe it, and ship it; finally you converge on an open-ended **Capstone** that breaks the single agent into a multi-agent team. Rather than lectures, you'll learn by *doing*: deploying models, building agents, grounding them in real data with a knowledge base, wiring action tools, shipping a hosted agent, and orchestrating a multi-agent system. Coaches guide discovery and unblock issues, but the learning is yours to own.\n\nBy the end, you'll have hands-on experience with the Microsoft Foundry platform, practical skills in prompt engineering and RAG (Retrieval-Augmented Generation), and a deployment-ready AI application to show for your work.\n\n---\n\n## Learning Outcomes\n\nBy the end of this hackathon, you will be able to:\n\n- Navigate Microsoft Foundry and provision an AI project with connected resources (`azd up`)\n- Deploy and compare models from the Microsoft Foundry model catalog\n- Build a named, versioned Foundry **agent** with a persona and guardrails\n- Ground the agent in your own data with an **Index + Foundry IQ knowledge base** (with citations)\n- Give the agent **action tools** via MCP that execute real operations\n- Evaluate outputs for quality and safety, including **red teaming**\n- Trace every answer end-to-end, then **deploy a hosted agent**\n\n---\n\n## Who is this for?\n\n### Students\n\nYou're a great fit for this hackathon if you:\n\n- Have some Python experience (variables, functions, pip) and understand REST APIs and JSON\n- Are curious about how AI models work and want to build with them\n- Have a GitHub account and access to Azure (via Azure Pass, Azure for Students, or a trial subscription)\n- Are ready to learn by solving real challenges, not watching tutorials\n\nNo prior Azure or AI experience needed. Just bring curiosity and a willingness to debug.\n\n### Coaches\n\nYou're ready to coach if you:\n\n- Are familiar with Microsoft Foundry model and agent concepts, and prompt engineering\n- Enjoy helping teams think through problems (instead of giving direct answers)\n- Can spend 6–8 hours supporting 2–3 teams\n- Have access to the [Coach Hub](docs/coach-hub.md) and solution materials in this repo\n\n---\n\n## Prerequisites\n\nBefore you start, make sure you have:\n\n- **Azure subscription**: Provided via Azure Pass, Azure for Students, or a free trial\n- **GitHub account**: With Codespaces access (required for the dev environment)\n- **Basic Python**: Comfortable with variables, functions, pip, and virtual environments\n- **Basic API knowledge**: Understand REST APIs, HTTP requests, and JSON\n- **VS Code familiarity**: Helpful, but not required (the devcontainer includes everything)\n\n---\n\n## Getting Started\n\n### 1. Open in GitHub Codespaces or Dev Container\n\nClick the badge below to open a fully configured development environment:\n\n[![Open in GitHub Codespaces](https://github.com/codespaces/badge.svg)](https://codespaces.new/olivomarco/ai-hackathon)\n\n**Alternative**: Open locally with [Dev Containers](https://code.visualstudio.com/docs/devcontainers/containers) in VS Code.\n\n### 2. Authenticate with Azure\n\nOnce your environment is ready, authenticate to Azure:\n\n```bash\naz login\n```\n\nFollow the prompts to sign in with your Azure account. This connects your workspace to your Azure subscription so you can provision Microsoft Foundry resources.\n\n### 3. Start with Foundations (or bootstrap)\n\n**Path A (beginner, recommended):** Read [Foundations](challenges/foundations/README.md) and work through Steps 1–4 to stand up a deployed, grounded Northfield IQ Assistant.\n\n**Path B (advanced skip):** Materialize the Foundations end-state with one bootstrap (~10–15 min), verify the single checkpoint, then jump straight to the Advanced tier:\n\n```bash\nazd up                                   # provision Foundry + AI Search + App Insights + ACR\n./scripts/setup-foundations.sh           # build the agent + index + IQ knowledge base\npython scripts/validate-foundations.py   # ✅ asserts the Foundations end-state\n```\n\n---\n\n## Challenges\n\nThe curriculum is a **tree of three tiers** — climb a guided trunk, fan out across modular branches,\nthen converge on an open summit. **Tier 1 Foundations** is one guided, linear challenge with four\nordered steps (everyone does it). **Tier 2 Advanced** challenges are modular and can be attempted in\n**any order** — they all assume the Foundations end-state. **Tier 3 Capstone** is an open-ended\ndesign brief that composes everything into a multi-agent system.\n\n```text\n  TIER 1  FOUNDATIONS (guided, linear, everyone)\n    Step1 --\u003e Step2 --\u003e Step3 --\u003e Step4  \u003c-- Foundations end-state\n           |\n           v\n  TIER 2  ADVANCED (modular, pick any order)\n    Action Tools | Evaluation+RedTeam | Tracing | Deploy\n    deepeners: Fabric IQ | Voice Live | Build a UI | Copilot-Assisted\n           |\n           v\n  TIER 3  CAPSTONE (open-ended design brief)\n    Northfield IQ multi-agent: triage/router fans out to specialists\n    (knowledge, actions), then converges.\n```\n\n### Tier 1 — Foundations (`challenges/foundations/`)\n\n| Step | Title | Duration | Difficulty | Builds toward end-state |\n|---|-----------|----------|------------|------------|\n| 1 | [Setup \u0026 Provisioning (Foundry + AI Search)](challenges/foundations/README.md#step-1--setup--provisioning-foundry--ai-search) | 30 min | ⭐ | Infra live; `.env` contract |\n| 2 | [Model Selection \u0026 the Playground](challenges/foundations/README.md#step-2--model-selection--the-playground) | 45 min | ⭐ | A chosen model + system instructions |\n| 3 | [Your First Agent](challenges/foundations/README.md#step-3--your-first-agent) | 45 min | ⭐⭐ | A named, versioned agent |\n| 4 | [Knowledge Base — Index + Foundry IQ](challenges/foundations/README.md#step-4--knowledge-base-index--foundry-iq---foundations-end-state) | 1.5 hr | ⭐⭐⭐ | **Grounded agent w/ citations (END-STATE)** |\n\n### Tier 2 — Advanced (modular · any order)\n\nEach Advanced challenge offers two paths: a **Guided** path (revised, honest time) and a longer\n**Build-from-scratch** path with fewer placeholders. Both are graded by the same `validate.py`.\n\n| Challenge | Guided | Build-from-scratch | Difficulty | Key Skills |\n|-----------|--------|--------------------|------------|------------|\n| [Action Tools — Make the Agent Do Work](challenges/advanced-action-tools/README.md) | ~45 min | ~1.5 hr | ⭐⭐⭐ | MCP tool, tool-approval loop |\n| [Evaluation \u0026 Red Teaming](challenges/advanced-evaluation-redteam/README.md) | ~1.25 hr | ~2 hr | ⭐⭐⭐⭐ | NLP metrics + adversarial safety |\n| [Tracing \u0026 Observability](challenges/advanced-tracing-observability/README.md) | ~1 hr | ~1.5 hr | ⭐⭐⭐⭐ | OTel GenAI → App Insights → KQL |\n| [Deploy as a Hosted Agent](challenges/advanced-deploy-hosted-agent/README.md) | ~60–90 min | ~1.5 hr | ⭐⭐⭐⭐⭐ | `azd ai agent`, hosted endpoint |\n\n**Extras** (optional, modular) — re-slotted by their role in the tree:\n- **Capstone-feeders**: Magentic Workflows, MAF + Hosted Long-Running Agents — the strongest content\n  feeds straight into the Tier 3 multi-agent build.\n- **Capstone companion**: Build a UI — a web front-end for your agent (or agent team).\n- **Deepeners**: Fabric IQ, Give It a Voice (Voice Live), Copilot-Assisted Build — extend one concept.\n\nSee the `challenges/extra-*` folders.\n\n### Tier 3 — Capstone (`challenges/capstone-multi-agent/`)\n\nThe open-ended summit: break the single Northfield IQ Assistant into a **multi-agent team** — a\n[triage/router that fans out to specialist agents and converges](challenges/capstone-multi-agent/README.md#the-agent-org-chart-role-as-agent),\norchestrated with the **Microsoft Agent Framework (MAF)**. It's a **design brief, not a placeholder-fill** —\nyou decide the org-chart and wire the graph.\n\n| Capstone | Time | Difficulty | Prereqs |\n|----------|------|------------|---------|\n| [Northfield IQ, the Team — Multi-Agent Orchestration](challenges/capstone-multi-agent/README.md) | 2–2.5 hr core (+1 hr Magentic stretch, +1.5 hr hosted variant) | ⭐⭐⭐⭐⭐ | Foundations end-state **+** Action Tools |\n\n**Make it your own:** the capstone is the best place to reskin — swap Northfield for your domain\n(insurance, factory ops, retail) and demo *your* agent team.\n\n**Total guided path (Foundations + all four Advanced): ~7.25 hours** + **~2.5 hr Capstone** — a clean\nmulti-day story. For a **1-day event**, run Foundations + 2–3 Advanced challenges and save the Capstone\nfor a second day or a follow-up sprint.\n\n---\n\n## Repository Structure\n\n```\nai-hackathon/\n├── README.md                          # ← You are here\n├── azure.yaml                         # azd project (golden-path provisioning)\n├── infra/                             # Bicep — Foundry + AI Search + App Insights + ACR\n├── scripts/                           # deploy.sh, setup-foundations.sh, validate-foundations.py, cleanup.sh\n│   └── action-backend/                # Action Tools REST API + FastMCP server (provided)\n├── challenges/                        # Challenge content and solutions\n│   ├── foundations/                   # Tier 1 — guided, Steps 1–4\n│   ├── advanced-action-tools/         # Tier 2 — modular, any order\n│   ├── advanced-evaluation-redteam/\n│   ├── advanced-tracing-observability/\n│   ├── advanced-deploy-hosted-agent/\n│   ├── capstone-multi-agent/          # Tier 3 — open-ended MAF capstone\n│   └── extra-*/                       # Tier 2 — Extras (optional)\n├── resources/sample-data/             # Northfield University FAQ corpus (knowledge base source)\n├── docs/                              # Supporting documentation (Jekyll/GitHub Pages)\n├── .devcontainer/                     # Dev environment config (Python, Azure CLI, azd)\n├── .github/                           # Copilot enablement (skills, copilot-instructions) + workflows\n├── .vscode/mcp.json                   # MCP servers: azure, foundry-mcp, microsoft-docs\n└── .env.sample                        # The .env variable contract (never commit a real .env)\n```\n\nEach challenge folder contains:\n- `README.md`: the challenge brief (what to build)\n- `solution.md`: the solution guide (coaches only)\n- Sample data or starter code (if needed)\n\n---\n\n## For Coaches\n\n### Getting Started\n\nVisit the [Coach Hub](docs/coach-hub.md) for tips on facilitating WTH events and working with your teams.\n\n### Solution Guides\n\nSolution guides for each challenge are included in this repo under `challenges/*/solution.md`. These are **for coaches only**; share judiciously to encourage discovery over answers.\n\nClone or access this repo locally and navigate to the challenge solution you need.\n\n### Quick-Start Coaching Checklist\n\n1. Verify all participants have Azure subscriptions and Codespaces access\n2. Review the challenge brief before your team starts\n3. Walk through Foundations Step 1 with them to confirm the environment works\n4. For each challenge, guide them toward the solution without giving it away\n5. Use the solution guide to unblock them if they're truly stuck\n\n---\n\n## Resources\n\n- **[Microsoft Foundry Documentation](https://learn.microsoft.com/azure/foundry/)**: Official docs and tutorials\n- **[Azure AI Learning Path](https://learn.microsoft.com/training/paths/develop-generative-ai-solutions-azure-ai-foundry/)**: Structured training modules\n- **[Microsoft AI Skills Navigator](https://microsoft.com/ai/skills)**: Browse AI and cloud certifications\n- **[What The Hack Format Guide](https://microsoft.github.io/WhatTheHack/)**: Learn about WTH events\n\n---\n\n**Ready to build?** Start with [Foundations](challenges/foundations/README.md).\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Folivomarco%2Fai-hackathon","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Folivomarco%2Fai-hackathon","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Folivomarco%2Fai-hackathon/lists"}