Catalog / AI-100
AI-100Emerging techAI & Emerging Tech
Enterprise AI adoption done properly: use-case triage, RAG architecture, security controls, and production readiness on the Microsoft stack.
About this program
Most enterprise AI initiatives stall between demo and production. This program trains architects to make AI adoption governable: triage use cases honestly, design retrieval-augmented generation on Microsoft Foundry (formerly Azure AI Foundry), integrate models behind secure serverless patterns, and put evaluation, cost control, and observability in place before the first production release.
Who this is for
- Architects asked to define their organization's AI adoption approach
- Senior engineers building AI-assisted features into enterprise systems
- Platform and security leads responsible for AI guardrails
What you will be able to do
- Triage AI use cases by value, feasibility, and risk before committing budget
- Design RAG architectures with grounded data, evaluation, and honest failure modes
- Apply security, privacy, and responsible-AI controls to model integrations
- Take an AI workload to production readiness with cost and observability discipline
Program modules
Delivered over 6 weeks of live tutor-led sessions, applied work, and structured review.
The enterprise AI landscape
- Model families, hosting options, and the build-vs-buy spectrum
- Use-case triage: value, feasibility, data readiness, risk
- Why pilots fail: the demo-to-production gap
- Governance stakeholders: security, legal, data, and the business
Platform foundations on Microsoft Foundry
- Microsoft Foundry (formerly Azure AI Foundry): projects, models, deployment options
- Model selection and endpoint patterns for enterprise use
- Prompt engineering as an engineering discipline
- Cost structure: tokens, throughput units, and budgeting
Retrieval-augmented generation architecture
- RAG end to end: ingestion, chunking, embedding, retrieval, synthesis
- Azure AI Search and vector retrieval design choices
- Grounding quality: evaluation sets and answer faithfulness
- When RAG is the wrong answer: fine-tuning and simpler alternatives
Security, privacy, and responsible AI
- Data boundaries: what leaves your tenant and what must not
- Identity, network isolation, and private endpoints for AI services
- Content safety, abuse monitoring, and human oversight
- Regulatory posture and auditability for AI features
Integration and serverless patterns
- Event-driven and API-fronted AI integration architectures
- Azure Functions and container patterns for model orchestration
- Resilience: timeouts, retries, fallbacks, and degradation paths
- Versioning models and prompts safely in production
Agentic workflows and production readiness
- Agent patterns: when autonomy is justified and when it is not
- Evaluation and regression testing for non-deterministic systems
- Observability: tracing, token cost, and quality dashboards
- Capstone: present a production-ready AI architecture for review
Hands-on components
- Guided Azure environment with Microsoft Foundry project access
- Build and evaluate a working RAG pipeline against provided datasets
- Capstone AI architecture review with structured critique
Ready to apply?
Apply any time. Admissions reviews your fit, confirms the next available cohort, and issues an invoice. Your seat is confirmed on payment.