7.5+ years building the systems 5,000+ employees run on: a multi-agent AI assistant on Microsoft Teams, a real-time integration backbone across 15+ systems, and an enterprise data lake on Fabric — leading ~25 engineers, shipping production systems in 3–60 day timelines.
Namakkal, Tamil Nadu, India · +91 94863 04676 · [Open to — set your availability line]
RoleArchitect & lead. Teams bot in C# / .NET 8 / Bot Framework SDK on Azure App Service; central Logic App router dispatching to 4 production specialist agents (Resource Info, Leave, ADO Sync, Knowledge Base) integrating Pulse, NetSuite, CRM and more. End-to-end observability in Application Insights; 14+ prompt iterations.
Under the hoodThe Knowledge Base agent is production RAG on Azure AI Search — 15+ country- and domain-specific policy indexes fronted by a query classification + rewriting layer that resolves follow-up questions into self-contained searches. Multilingual via Azure AI Translator.
Next generationArchitected 3 next-gen agent systems — L&D Agent, PRM Agent (10 sub-agents), Opportunity Agent (7 sub-agents) — and established reusable patterns: promise-ledger, action registry, domain-driven sub-agent decomposition.
Public write-up: building & deploying a Teams bot on Azure, end to end →
RoleSole architect. Real-time integration layer connecting 15+ internal and SaaS systems across HR, Sales, Finance and Resource Management — 50+ Logic App workflows behind API Management, including Greenhouse ATS at 200+ job requisitions/month. Eliminated operational silos organization-wide.
The hard 20%OAuth token-invalidation race conditions, WAF false positives on legitimate traffic, schema-validation overflows, private ILB triggers with APIM as the sole front door, fixed egress via NAT gateway for partner whitelisting.
RoleArchitect & team lead (4 data engineers + 3 Power BI analysts); migrated 2 source systems personally. Medallion architecture, 10+ curated Gold datasets.
The interesting partA dynamic GraphQL API as an MCP layer — the AI model reads the schema, constructs its own queries, and answers conversationally. Democratized data access for Operations, FP&A, HR, Sales and the C-suite; HR Data Agent demoed to the CEO.
WhatA knowledge graph built entirely from the code layer — table relationships, per-column definitions, and join paths — designed to power natural-language query answering and automated report generation. The ontology layer between the warehouse and the agents, so schema context stops being hand-fed into every prompt.
Public write-up: a knowledge graph for a warehouse with no foreign keys →
ScopeDriving the security architecture across the estate: least-privilege integration identities (after auditing an over-broad access grant), bulk secret centralization into Key Vault with managed identities, private ILB workflow triggers with API Management as the only public surface, Front Door / WAF rule tuning, and secret-rotation discipline.
NowBuilding a per-application security assessment framework — methodology and dimensions first, piloted application-by-application before any tooling.
Public write-up: security isn't a checklist — it's a quarterly habit →
RoleArchitected the full greenfield solution (Angular + .NET + Fabric Gold) and directed the team: catalog UI, dynamic filters, favorites, multi-format export (Excel, CSV, PDF, JSON). Single reporting hub replacing fragmented reporting for FP&A, HR, Operations, Sales and executive leadership.
RoleDesigned a two-tier parallel architecture ingesting Teams, Email and CRM data into Fabric Gold for Copilot insights. The COO uses it for deal tracking across ~15 key accounts.
RoleIndependently designed the complete schema and configurable workflow engine; led the full lifecycle to production. Live within 60 days, now used organization-wide across all geographies.
RoleSolo architect — schema, workflow, competency framework, reporting. Enterprise-wide skill visibility, delivered end-to-end in 45 days.
RoleDesigned allocation workflows with 38 test scenarios and 7 compliance checks. Led the EU rollout in Serbia — 2 months, 2,000+ employees, 20+ project managers — unifying global operations under one framework.
RoleConsolidated approval workflows into a single auditable platform, and replaced ~90 on-prem SSRS reports with SheetJS Excel builders — removing the critical blocker for the Azure cloud migration.
RoleJMeter load testing across core platform screens and integration paths (P95 backend 0.72s), read/write contention analysis, and the capacity plan — tier upgrade, load balancing, autoscaling — with monthly cost forecasting to leadership.
Own architecture and delivery for enterprise platforms serving 5,000+ employees. Scope grew from a single platform to the full internal estate — applications, the integration backbone, the data lake, and the AI agent ecosystem — plus leadership of ~25 engineers. The title stayed the same. The blast radius didn't.
Enterprise web services and RESTful APIs for examination management and certification workflows — full lifecycle from requirements through integration testing to production deployment.
An EEE graduate who ended up architecting AI platforms — the circuit-analysis instinct never left.
I'm looking for a role where the title matches the scope. A short call is the fastest way to know — reach out.