Cloud and AI engineer at Hexaware Technologies, two years in — building automation, enterprise Copilots, and GenAI agents on Azure.
My work follows one pattern: something is being done by hand — a migration, a permissions audit, a report rebuilt every Monday, a ticket queue at 2 a.m. — and it shouldn't be. I find it, script it in PowerShell or Python, and these days I wire it through an AI agent instead of only a cron job.
At work that has meant migrating 5,000+ users of on-premises OneDrive data to Microsoft 365 with ShareGate and PowerShell (overnight cutovers running unsupervised); deploying StorageX Analytics against 9 storage VMs across 2 clusters into 10 Power BI reports the client formally thanked us for; migrating 132 Power Apps and 50+ Power Automate flows across 4 tenants; building Transcend, a Power App with a Copilot Studio agent that now runs 20+ projects; and delivering a Terraform landing zone POC for a client evaluation. On the same engagements I deployed two production Copilot Studio agents — an L1 support agent and a PMO assistant — grounded in internal SharePoint FAQs, wired to ticketing, live across all four tenants.
Outside work I build to go deeper: InfraGenie, an AIOps platform for Azure that uses a local LLM (Qwen2.5-Coder on Ollama in Docker) with structured prompts and self-remediation loops; a two-region DR environment in Terraform; and labs on AKS GitOps, policy enforcement, and observability.
Currently doubling down on GenAI engineering: agentic workflows with LangGraph, Model Context Protocol (MCP), RAG on Azure AI Search, LLMOps, and deploying AI workloads on AKS. Working toward AZ-104 (Oct 2026), then AZ-400 and AI-102.
📍 Gondia, Maharashtra · 🏢 Hexaware, Chennai · 🌏 Open to GenAI / AI Platform / Cloud roles · ⏱ IST (UTC+5:30)
|
A self-service platform for Azure infrastructure. Engineers describe what they need in plain English; the platform matches the request to a reusable Terraform module and runs policy checks before anything is applied. After deployment it keeps watching — it remediates common failures on its own and logs a ServiceNow ticket. A reporting agent publishes 10+ operational reports on FinOps spend, orphaned VMs, and drift.
|
A two-region, three-tier Azure environment — if the primary region fails, how fast can the app come back? Provisioned entirely in Terraform across Central and South India. VM Scale Sets serve web/app tiers, geo-replicated Azure SQL holds data, Traffic Manager + Load Balancers route traffic.
|
|
A Kubernetes lab for cost and governance: Kubecost and Prometheus track spend and utilization; manifests are GitOps-managed and validated on a KinD loop before reaching a real cluster. Reusable Terraform modules with platform guardrails for repeatable AKS delivery.
|
A Kubernetes policy and observability lab: Kyverno admission policies enforce image and resource rules; Grafana dashboards over cluster metrics and traces. Built to understand what "secure by default" actually costs to run day-to-day.
|
| Azure |       |
| IaC & DevOps |        |
| AI & Copilots |     |
| Observability |    -1B3A6B?style=flat-square&logo=microsoftazure&logoColor=white){kind=icon} |
| Scripting |      |
| Microsoft 365 |      |
| OS & Networking |    |
| Certification | Issuer | Status |
|---|---|---|
| Azure Fundamentals (AZ-900) | Microsoft | ✅ Mar 2025 |
| Azure Administrator Associate (AZ-104) | Microsoft | 🟡 Targeting Oct 2026 |
| DevOps Engineer Expert (AZ-400) | Microsoft | 📅 Planned |
| Terraform Associate (004) | HashiCorp | 🔄 In progress |
| Azure AI Engineer Associate (AI-102) | Microsoft | 📅 Planned |
- GenAI & Agents deep-dive — LangGraph agents, MCP servers, RAG on Azure AI Search, LLMOps observability. Deploying agent workloads on AKS.
- AZ-104 — Azure Administrator (this month).
- Python + FastAPI — production-grade async APIs for AI services.
- InfraGenie v2 — adding MCP tools, LangGraph multi-agent architecture, and AKS deployment.
- Production Kubernetes depth — cluster operations, GitOps, cost governance, troubleshooting.
Automate the manual parts. Build agents that scale.