I build the tooling that makes AI-assisted engineering measurable, affordable and safe to adopt at enterprise scale.
Most of my work lives at the intersection of three things: agentic developer tooling (VS Code extensions and orchestration built on the GitHub Copilot SDK), AI cost engineering (token compression, AI-credit modelling, billing observability), and cloud economics (BOQ generation, multi-cloud TCO, competitive bid strategy for Azure vs. AWS/GCP).
role: Cloud & AI Solution Engineer @ Microsoft
based_in: Bhopal, India
focus: [ Agentic DevTools, AI Cost Engineering, Cloud FinOps, Azure + Kubernetes ]
building: [ VS Code extensions, Copilot SDK agents, benchmark harnesses, enablement labs ]
ask_me_about:
- Squeezing 40%+ out of your Copilot token bill without losing answer quality
- Designing defensible Azure BOMs & BOQs for competitive RFPs
- Shipping CI/CD to AKS with Azure Pipelines, ACR and agentic DevOps
open_to: [ Developer tooling, applied GenAI, IoT & green tech collaborations ]
|
|
|
|
|
|
AI Cost & Token Engineering — making Copilot economics predictable
| Project | What it does |
|---|---|
| stenographer-mode | Shorthand-first prompt compression with VS Code prompt bundles, exact token benchmarking and cross-platform starter packs |
| token-optimization-simulator | Interactive AI-credit optimisation simulator for VS Code and the Copilot CLI |
| aic-lens | VS Code extension that estimates Copilot Chat usage live from your prompt + context files, before you spend it |
| copilot-billing-observatory | Enterprise Cloud billing toolkit — API reports, local dashboard, CLI automation and scheduled Outlook delivery |
| ai-credit-calculator | AI-credit modelling for usage-based billing readiness |
| token-optimisation | 10 hands-on exercises + a "Future You" simulator app, built for usage-based-billing readiness |
Agentic Developer Tooling — built on the GitHub Copilot SDK
| Project | What it does |
|---|---|
| github-pixel-squad-flow | Chat-first multi-agent orchestration in VS Code, with a hybrid runtime spanning Copilot and Claude language models |
| Guardrail | Real-time code guardrails — LLM-powered security, compliance and best-practice analysis as you type |
| What-Changed | Agent change tracker: captures every AI file edit with the prompt, reasoning and conversation context behind it |
| sunsetwatch | Multi-ecosystem dependency sunset scanner (npm, pip, Cargo, Go) — web app and CLI |
| se-brain | LLM-powered personal wiki that turns scattered research into a citation-grounded, interlinked knowledge base |
| sesp-session-planner | VS Code extension for planning hackathons, labs and challenges across the Azure + GitHub suite |
| multi-agent-developer | Multi-agent persona workspace for custom Copilot agents |
| mypromptdaily | CLI for running and scheduling daily M365 Copilot prompts through Outlook desktop |
Cloud Economics, Compete & Benchmarking
| Project | What it does |
|---|---|
| CloudQuote | Provider-extensible cloud BOQ generation and cost comparison, driven from Copilot |
| cloud-rfp-bom-skill | Copilot agent skill for defensible Azure-first RFP responses, BOMs, BOQs and win strategy vs. AWS & GCP |
| jev-vs-llm-benchmark | Provider-agnostic benchmark harness — quality, latency, tokens, routing and case-level evidence |
| jev-vs-foundry-benchmark | Local-first variant of the same harness for accuracy and routing evaluation |
Enablement, Workshops & Labs — used to train engineers at scale
| Project | What it does |
|---|---|
| github-copilot-labs | Self-paced labs across Spring Boot, React, Python, .NET, modernisation and Figma MCP |
| azure-devops-acr-aks-workshop | Participant-ready Azure Pipelines workshop: Node.js tests → ACR image builds → AKS deploys |
| Azure-DevOps-X-AKS | End-to-end agentic DevOps workshop, Azure DevOps to AKS |
| ghcp-pm-spec-kit | BRD → full project artefacts in one pipeline, using spec-kit-driven development |
| GuardRails-and-Secure-Coding | Secure coding lab modules for Copilot |
| github-copilot-flutter-demo-kit | Canned demo kit: business requirements → built and tested Flutter app |
| MCP-OS-Ticket-Lab | Hands-on Model Context Protocol lab |
| Governance-Hackathon | NeGD hackathon materials, data and challenge guides |
Cloud & Infrastructure
AI & Agents
Languages
Frameworks & Data
Cards are generated from the GitHub API and committed to this repo by a scheduled workflow — no third-party card service, so they never rate-limit or go offline. Language shares cap each repository's contribution so vendored dependencies don't skew the result.
