Azure AI Search, Functions, Container Apps, Key Vault, and Entra ID

Overview

• Azure AI Search provides searchable indexes and retrieval capabilities.
• Functions runs event-driven code; Container Apps runs containerized services/jobs.
• Key Vault manages secrets/keys, while Entra ID provides identity capabilities and access integration.

Why this matters

A production AI system needs retrieval, compute, credentials, and identity in addition to inference.

An everyday example

A company needs an archive, offices, locked key storage, and employee passes—not just a writing assistant.

A technical example

Container Apps hosts Node/FastAPI services; Functions reacts to document uploads; Search indexes policy chunks; managed identities access approved resources.

How it works

1. Authenticate user/service
2. Receive task
3. Store/queue input
4. Execute compute
5. Retrieve via Search
6. Call model
7. Write authorized results and telemetry.

Architecture and flow

Entra-authenticated app → Functions/Container Apps → Search + business services + model, with secrets controlled through Key Vault.

The insurance assistant example

Document-upload events trigger parsing and indexing while the web API remains responsive.

When to use it

An Azure-centered stack with workloads fitting the chosen service limits and operating model.

When to avoid it

Do not run long GPU work in a short-lived function by assumption; verify supported hosting, duration, and hardware.

Interview answer

“I separate identity, secrets, compute, search, and model access. Each service has one clear role and least-privilege access.”

Benefits, trade-offs, and production considerations

• Private connectivity, scale-to-zero/cold starts, index capacity, and permission management affect latency and reliability.
• Managed identity reduces static secrets but does not automatically enforce customer-level authorization.

Official references: AI Search, Container Apps, Key Vault, Entra, and Durable Functions.