I'm a Senior Backend & AI Engineer from India with 12+ years of hands-on experience building production systems that serve real users. I specialize in AI/LLM integration, scalable APIs, and real-time systems.
- 🔭 Currently building an AI-powered consumer insights platform — LLM pipelines, RAG, face redaction, real-time streaming
- 🤖 Working daily with OpenAI GPT-4o, Azure AI, YOLO, Whisper, LangChain in production
- 🔌 Building MCP servers to connect AI agents (Claude, Cursor) to live databases
- 🏗️ Architecting backend systems with Go + DDD/CQRS for high-stakes services
- 💳 Built SaaS products end-to-end — Stripe, SSO, multi-tenant, subscription lifecycle
- 🏨 Shipped an IoT platform used by hotels worldwide for keyless entry
- 🌐 Looking for remote contract/part-time work with US or India-based companies (Bangalore, Chennai, Hyderabad, remote)
- ⏰ IST evenings = perfect overlap with EST/PST business hours
Backend & Languages
AI / ML / LLMs
Databases
Cloud & DevOps
Integrations & Tools
| Domain | What I Did | Tech Used |
|---|---|---|
| 🤖 AI/LLM Pipelines | Vector DB + embeddings for auto-summarization of 10K+ consumer videos | OpenAI, Azure, pgvector, BullMQ |
| 🔌 MCP Servers | AI agent ↔ database bridge with SQL safety layer for Claude & Cursor | TypeScript, PostgreSQL, MySQL, MongoDB |
| 👤 Computer Vision | Privacy-compliant face redaction at enterprise scale | YOLO, Azure Video Indexer |
| 🎙️ Speech-to-Text | Transcription + keyword extraction from audio datasets | Whisper, VoiceBase |
| 📡 Real-Time Streaming | Live AI-moderated sessions with GPT-4o responses | Agora, LiveKit, WebSockets |
| 🏗️ Go / DDD Backend | Contract management with FSM, CQRS, aggregate roots | Go, PostgreSQL, Chi, JWT |
| 🏨 IoT Platform | Keyless hotel entry — encrypted mobile-to-lock communication | Node.js, LDAP, Electron |
| 💳 SaaS & Payments | Multi-product subscription platform with Stripe | Stripe, SSO, Node.js |
| 🔐 Security Operations | SOC portal with QRadar integration via VPN | Node.js, Cron, CMDB |
NestJS microservice that ingests customer feedback (API, webhooks, bulk CSV), generates OpenAI embeddings, and runs AI-powered analysis — semantic search, theme extraction, sentiment scoring, and summarization. TimescaleDB + pgvector for vector similarity search, BullMQ + Redis for async job processing.
Features: RAG pipeline with pgvector cosine search • HMAC-verified webhooks • Async embedding jobs via BullMQ • Sentiment + emotion detection • Theme & keyword extraction • GraphQL + REST API • Swagger docs at /api/docs • Docker ready
Production-ready MCP (Model Context Protocol) server that lets AI agents (Claude, Cursor, GPT) securely query databases. Read-only by default, SQL injection detection, supports PostgreSQL, MySQL & MongoDB. Plug into Claude Desktop or Cursor IDE in minutes.
Features: 5 MCP tools — query, schema, describe, sample data, table stats • SQL injection detection + write protection • Output as markdown, JSON, or CSV • Works with Claude Desktop, Cursor IDE, Claude Code CLI • Docker setup with ecommerce seed data • Query timeout + row limits
Production-grade Contract Management System in Go — Domain-Driven Design, CQRS, Clean Architecture. Full contract lifecycle FSM (Draft → PendingReview → Approved → Active → Expired/Terminated), PostgreSQL, JWT auth, GitHub Actions CI.
Features: DDD aggregate roots + value objects (Money, Clause, DateRange) • CQRS-lite command/query handlers • Contract FSM with domain events • Unit of Work + optimistic concurrency • Repository pattern with sqlx • Multi-stage Docker build + 6 DB migrations
Upload PDF/DOCX/TXT and chat with your documents — RAG pipeline with vector embeddings, source citations, conversation memory, and real-time streaming via WebSocket. Built-in chat UI and mock mode (no API key needed for demos).
Features: RAG with vector similarity search • Smart chunking with overlap • Conversation memory across sessions • Source citations per answer • Token-aware prompt management • Mock mode with TF-IDF (no API key needed) • Dark-theme chat UI included
Production-ready REST API that analyzes customer feedback using OpenAI GPT-4o — extracts summaries, keywords, sentiment & themes. Batch processing (50/request), JWT auth, rate limiting, analytics dashboard, CSV export, and webhook delivery. Docker ready.
Features: GPT-4o sentiment + keyword + theme analysis • Aspect-level sentiment breakdown • Batch processing (up to 50/request) • CSV export with filters • Webhook delivery to Slack/Zapier • Analytics aggregation API • Mock mode (no API key needed) • Rate limiting + JWT auth
Scalable WebSocket-based chat backend — rooms, typing indicators, message persistence, read receipts, online presence tracking, and a built-in test client. Node.js + Socket.io + MongoDB + Redis + Docker.
I'm actively looking for remote contract or part-time roles with US or India-based companies (Bangalore, Chennai, Hyderabad, or fully remote). If you need a backend engineer who can:
- ✅ Integrate AI/LLMs into your product (OpenAI, Azure, RAG, embeddings, MCP)
- ✅ Build scalable REST/GraphQL APIs and microservices in Node.js or Go
- ✅ Set up real-time systems (WebSockets, streaming, live pipelines)
- ✅ Design vector search & AI data pipelines for production workloads
- ✅ Work independently with minimal supervision across time zones
→ Let's talk: rohitkapoorfriend@gmail.com | LinkedIn