Languages
Backend
Data
Distributed / Realtime
Infrastructure
Testing / Observability
Frontend

Event-driven ecommerce platform built around transactional correctness — asynchronous processing, concurrency-safe inventory, payments, and search across a typed service boundary.

EVENT-DRIVEN  IDEMPOTENCY  CONCURRENCY

TypeScript · Fastify · PostgreSQL · Redis · BullMQ · Docker

Repository →

_{4-service TypeScript monorepo · typed domain events driving 8+ idempotent BullMQ queues with retry/dead-letter handling · Redis distributed locks + PostgreSQL conditional transactions for oversell-safe inventory · JWT with rotating refresh tokens, RBAC · Stripe Checkout with signed webhooks}

Production-oriented subscription platform combining realtime communication, asynchronous billing, and operational tooling across four independently scalable runtimes.

REALTIME  PUB/SUB  WEBHOOKS

Next.js · Redis · Socket.IO · BullMQ · Stripe · Razorpay

Repository →

_{Standalone Socket.IO runtime with authenticated rooms, Redis Adapter, and Redis Pub/Sub for horizontal scale · BullMQ workers for billing, draws, analytics, and leaderboards · Stripe/Razorpay webhook processing with HMAC verification and idempotent, replay-safe reconciliation · Prometheus + Grafana observability}

AI-powered study platform built around reliable structured generation, transactional persistence, and production-oriented observability.

AI GENERATION  TRANSACTIONS  OBSERVABILITY

Next.js · PostgreSQL · Gemini 2.5 Flash · Vitest · Prometheus

Repository →

_{Parallel transcript/title fetching with direct-video Gemini fallback · Schema-enforced JSON generation with Idempotency-Key handling, stale-request reclaim, and atomic PostgreSQL persistence · Structured JSON logging, request IDs, Prometheus metrics, health checks, TLS-verified DB connections, GitHub Actions CI}

_{stats.svg and top-langs.svg are static assets generated and committed by this repository's own GitHub Actions workflow — see SETUP.md for how it works.}

IEEE PUBLISHED

Multimodal Deep Learning for Chest X-Ray Disease Classification

IEEE-published research on multimodal CNN/MLP fusion for chest disease classification and model explainability.

PyTorch · CNN/MLP · Multimodal Learning · Grad-CAM

Publication → _{(replace with the IEEE DOI/Xplore link)}

Distributed Systems · System Design · Backend Performance · Reliability Engineering