Skip to main content

All Projects

Systems and tools I've built or led — from validation engines to developer utilities.

VKYC AI Platform

End-to-end Video KYC platform with 4 polyglot microservices (NestJS, Go, Python, FastAPI) and 4 frontend apps. Features a 19-state lifecycle engine, WebRTC SFU via Mediasoup, 7-layer fraud shield, live Whisper STT co-pilot, and 98.7% OCR accuracy on Indian documents.

📈 Automated 80% of KYC verifications

NestJSGoPythonWebRTCDeepFaceRedis Streams

Scalet

Real-time team communication platform built with a Go backend and React 18 frontend. Features WebSocket hub/client model with clustering, workspace isolation at SQL level, glassmorphism UI with Framer Motion physics, and a pluggable preference engine.

📈 Designed for thousands, scales to millions

GoReactWebSocketPostgreSQLRedis

Simulation

Universal dynamic mock service platform. Smart policy builder with conditional routing (regex, JSONPath), response latency simulation, stateful sequences, transparent proxy for unmocked routes, and real-time analytics. React bundled into a unified Spring Boot JAR.

📈 Eliminated external mock dependencies

Java 17Spring BootReactAWS SQSMonaco Editor

Ultimate Finance Calculator

Full-stack portfolio analytics with AWS Lambda for daily NAV sync, S3 data lake for historical fund data, Supabase transaction ledger, and CloudFront CDN. Features XIRR calculation, interactive charts with Recharts, and PDF report generation.

📈 Serverless, zero-maintenance architecture

React 19AWS LambdaTerraformSupabaseTypeScript

Lets Settle

Full-stack settlement platform built with Next.js 16 and React 19. Docker Compose orchestration with PostgreSQL 17 and MinIO S3-compatible storage for payment workflow management and agreement tracking.

📈 End-to-end settlement workflow

Next.jsReact 19TypeScriptPostgreSQLDocker

NetworkNT Validation Engine

High-throughput JSON Schema validation engine for FinTech. Extended with custom keywords for Indian financial domain logic (PAN, Aadhaar, IFSC). Schema compilation cache dropped P95 from 45ms to 8ms.

📈 3x faster validation, 98% accuracy

JavaJSON SchemaSpring Boot

Go Validation Engine

Mission-critical validation service rewritten in Go for ultra-low latency. Goroutines for concurrent sub-schema validation, reducing infrastructure costs and P99 latency significantly.

📈 40% reduction in P99 latency

GogRPCDocker

Catalog Service

Java-based microservice for enterprise product catalogues. Clean architecture with domain-driven design, supporting high-concurrency reads via caching layers.

📈 Sub-50ms P95 response time

JavaSpring BootRedis