Software engineer at Hexaware Technologies, two years in, working across Azure, Microsoft 365 and automation. Most of my work follows one pattern: something is being done by hand, and it shouldn't be. I find it, script it, and put the output where people can see it.
At work that has meant migrating 5,000+ users to Microsoft 365 with ShareGate and 5 PowerShell scripts; deploying StorageX Analytics against a client's NetApp ONTAP estate and scripting collection across 9 storage VMs / 2 clusters into 10 scheduled Power BI reports; moving 132 Power Apps and 50+ Power Automate flows to a new tenant; and building Transcend, a Power App with a Copilot Studio agent that is now the primary tracking tool for our internal PMO team.
Outside work I build infrastructure projects to go deeper — InfraGenie, a two-region Azure DR environment in Terraform, and KubeScale-Ops. Being straight about it: my Terraform and Kubernetes depth comes from those projects and one client POC rather than years of production ownership. AZ-400 next, then CKA.
|
Turns a plain-English request into a reviewable Terraform plan — mapped to vetted modules, with policy checks and a cost estimate attached to the diff. Nothing reaches the subscription until a human approves the plan. Runs on AKS, with Argo CD for drift detection and Kubecost for node-level cost visibility. State is held remotely in Azure Storage. Architecture and infrastructure design are mine; application code was generated with AI coding agents under structured prompting. repo · HF model (fine-tuned Qwen2.5-Coder-7B, SFT) Terraform AKS Argo CD Kubecost Python Qwen2.5-Coder-7B
|
Active–passive two-region disaster recovery defined entirely in Terraform and deployed through GitHub Actions: Traffic Manager routing, auto-scaling VM Scale Sets and geo-replicated storage. Failed over end to end, then documented — including what broke on the way. Destroy-and-rebuild reproducible. Terraform Traffic Manager VMSS Azure SQL GitHub Actions
|
|
FinOps control plane where Kubecost + Prometheus track spend and utilization. Kubecost surfaced ~20% reclaimable spend from rightsizing in the test cluster; manifests are GitOps-managed and validated on a KinD loop before they reach a real cluster. Kubecost Prometheus Helm KinD GitOps
|
Kubernetes policy and observability lab: Kyverno admission policies enforcing image and resource rules, Istio for service-to-service traffic, and Grafana dashboards over the cluster. Built to understand what "secure by default" actually costs to run. Kyverno Istio Grafana Kubernetes
|
| Category | Technologies |
|---|---|
| Cloud |    |
| Containers |     |
| IaC & Automation |       |
| CI/CD & GitOps |     |
| Observability & FinOps |      |
| Microsoft 365 & Power Platform |       |
| Languages & APIs |     |
| AI / GenAI |    |
- AZ-400 (DevOps Engineer Expert) — sitting Sep 2026.
- CKA (Certified Kubernetes Administrator) — starting after AZ-400.
- NVIDIA NCA (AI Infrastructure Operations) — in progress.
- InfraGenie hardening — tighter policy checks and a clearer plan-diff review step.
- Microsoft Certified: Azure Administrator Associate (AZ-104) — Aug 2026 ✅
- Microsoft Certified: DevOps Engineer Expert (AZ-400) — expected Sep 2026
- NVIDIA NCA — AI Infrastructure Operations — in progress
- Claude Certified Associate — Anthropic — expected Sep 2026
