From AI demo to Azure deployment
A practical guide to building a secure, keyless, and observable Azure AI agent POC with Microsoft Foundry, Azure AI Search, Container Apps, Terraform, and Azure DevOps.
CloudSci field notes
AI engineering,
under production constraints.
One practical article each week about building AI systems and cloud platforms that remain secure, observable, reproducible, and useful after the demo ends.
Editorial pillars
Four systems, one production surface.
The blog follows the places where AI applications meet identity, infrastructure, reliability, and human decision-making.
AI Engineering
Evaluation, agents, retrieval, model operations, safety boundaries, and human review.
Cloud Platforms
Landing zones, identity, network boundaries, governance, FinOps, and resilient foundations.
Kubernetes & IaC
AKS, workload identity, Terraform modules, GitOps, policy, and repeatable delivery.
Production Practice
Observability, incident learning, architecture trade-offs, security, and operational readiness.
Written from experience
By Gowrisha C. Vishwa Kumar.
Senior Cloud Platform Engineer writing about the practical boundary between AI applications and the cloud platforms expected to run them safely.