I design and build backends that stay fast and reliable at scale, and I've led backend teams through the big architecture decisions.
- 🛒 Foodi: re-architected inventory into a distributed real-time system (90% fewer stock discrepancies), monolith → RPC microservice (2.8s → 1.9s)
- ☁️ Actionboard.ai (US SaaS): usage-metering & billing platform in Go, plus an async ML streaming pipeline on AWS with sub-100ms latency
- 🎓 Shikho: AI agent pipeline that turns lectures into quizzes and summaries, and payments for 2.5M+ users
- 🗺️ Mapnest: 600K+ geospatial records at sub-100ms p99 geocoding
- 📦 Distributed order platform handling 10,000+ orders at peak (message broker + RPC)
- 🎓 BSc in Computer Science & Engineering, RUET
| Project | What it does | Stack |
|---|---|---|
| stripe-usage-billing | Usage-based (metered) billing on Stripe with verified, idempotent webhooks | Go · Postgres · Stripe |
| ai-booking-agent | AI receptionist that books appointments via tool calling, with no double-booking | TypeScript · Claude/OpenAI |
| rag-docs-chatbot | Chatbot that answers from your documents with citations and an honest "I don't know" | Python · FastAPI |
| mcp-database-server | MCP server that lets Claude or ChatGPT safely query a company database (read-only) | TypeScript · MCP · SQL |
| n8n-ai-automation-pack | Production-grade AI workflows: email triage, lead scoring, invoice extraction | n8n · OpenAI |
- Backend architecture: microservices, queues, caching, database design, scaling and rescues
- SaaS backends: auth, multi-tenancy, subscriptions and billing, APIs
- AI agents & chatbots: RAG on your docs, tool-using agents, MCP servers, LLM features in your app
- Automation: n8n, Make and Zapier workflows with AI, built like real software
- Fix & finish: slow APIs, failing jobs, and apps built with Lovable, Bolt or Cursor
Stack: Go · TypeScript/Node.js · Python · PostgreSQL · Redis · RabbitMQ · Elasticsearch · gRPC · AWS (Lambda, SQS, ECS) · Docker · Kubernetes
📩 Have a project? Message me on Fiverr. I reply fast.
