Backend Software Engineer | Distributed Systems | AI Systems | Cloud-Native
Building scalable backends, real-time systems, and open-source developer tools
Software Engineer at GeekyAnts with 2+ years of experience building production-grade distributed systems, real-time streaming platforms, and AI-driven architectures. I design systems that handle 50K+ MAU with 99.8% uptime and sub-200ms response times.
Geekathon 2026 Winner — Built Moshai, an autonomous multi-agent fraud detection system processing transactions in <200ms with explainable AI decisions.
| System | Scale | Key Tech |
|---|---|---|
| Live Debate & Streaming Platform | 5K+ concurrent users, <100ms latency | NestJS, WebSockets, Redis Pub/Sub, Elasticsearch KNN |
| Video Competition Platform | 5K+ MAU, 1K+ videos/day | BullMQ, AWS Rekognition/Transcribe/Bedrock, FFmpeg |
| Travel Marketplace | $50K+ monthly revenue | Stripe subscriptions, idempotent webhooks |
| Fintech Loan Platform | Secure auth system | JWT, refresh tokens, RBAC, MFA |
|
Dev-Crew 24 AI agents for code review, testing, security, DevOps — one CLI |
SQL Optimizer AI-enhanced PostgreSQL query optimization & performance analysis |
npm install -g dev-crew # 24 AI agents in your terminal
npm install -g sql-optimizer # PostgreSQL query optimizationBackend NestJS, Express, REST APIs, GraphQL, Microservices
Databases PostgreSQL (advanced), Prisma, MongoDB, Redis, Elasticsearch
AI/ML RAG, Vector Search, XGBoost, Sentence Transformers, SHAP
Real-Time WebSockets, Socket.IO, SSE, Redis Pub/Sub, BullMQ
Cloud AWS Lambda, S3, CloudFront, Rekognition, Transcribe, Bedrock
Video FFmpeg, Adaptive Bitrate Streaming, Multi-Resolution Encoding
Payments Stripe API, Subscriptions, Webhooks, Idempotency
Observability Prometheus, Grafana, APM, p95/p99 Metrics
DevOps Docker, CI/CD, Nginx, DigitalOcean, GitHub Actions
- Reduced API latency from 800ms to 120ms across production systems
- Designed hybrid semantic + lexical search improving content discovery accuracy by 85%
- Built AI moderation pipeline automatically flagging 98% of inappropriate content
- Optimized PostgreSQL queries from 2s to 150ms (p95)
- Processed 87+ transaction signals for real-time fraud detection at P95: 85ms
- 2x GeekWiz Award recipient for technical excellence
Dev-Crew · SQL Optimizer · npm · LinkedIn

