Capstone
The capstone challenge brings together skills from all six domains into a single end-to-end AI solution. This mirrors what you'll face in production and on the exam — questions that span multiple services and require architectural thinking.
You'll build a complete solution that provisions resources securely, implements a RAG pipeline with Azure OpenAI and AI Search, processes documents with Document Intelligence, adds NLP enrichment, and deploys with monitoring and responsible AI practices. This is your final validation before taking the exam.
Treat this as your dress rehearsal. If you can complete the capstone without referring back to earlier challenges, you're ready for the AI-102.
What You'll Learn
- Integrate multiple Azure AI services into a cohesive solution
- Make architectural decisions (which service for which task)
- Implement production patterns (error handling, retry, monitoring)
- Apply responsible AI practices across the entire pipeline
- Deploy and validate an end-to-end AI application
Skills Measured
This challenge touches skills from all six domains:
- Resource provisioning and security (Domain 1)
- RAG and prompt engineering (Domain 2)
- Tool use patterns (Domain 3)
- Image/document processing (Domains 4 & 6)
- Text analytics and language understanding (Domain 5)
- Search indexing and querying (Domain 6)
Challenges
| # | Title | Key Topics |
|---|---|---|
| 49 | End-to-End AI Solution | Multi-service integration, production patterns, full pipeline |
Prerequisites
- Completed all previous domains (Challenges 01–48)
- All Azure resources provisioned (AI Services, OpenAI, AI Search)
- Confidence with both Python/C# SDKs and REST APIs
- Understanding of production deployment patterns