MS in Computer Science @ Stevens Institute of Technology | Hoboken, NJ
- 🔬 Building a Distributed ML Profiling Platform @ Stevens — energy metrics from 100+ edge devices in real-time
- 🚀 Former Founding Engineer — scaled a placement platform to 500+ students, 10K+ API requests/day
- 🧠 Passionate about Distributed Systems, System Design, and AI/ML
- 💬 Ask me about Java, Python, gRPC, Redis, TensorFlow, Distributed Architectures
- 🌐 Portfolio: kunjpandya.com
- 📫 Reach me at kunjspandya@gmail.com
🔬 Backend Engineer — Distributed ML Platform Stevens Institute of Technology · Nov 2025 – Present · Hoboken, NJ
- Architecting a distributed profiling platform for federated ML collecting energy metrics from 100+ edge devices in real-time
- Built scalable backend with TensorFlow.js, Socket.IO & Redis for per-layer power data across heterogeneous device networks
🚀 Founding Engineer — Placement Intelligence Platform D.J. Sanghvi College of Engineering · May 2024 – Jun 2025 · Mumbai, India
- Built & scaled backend from scratch for 500+ students, handling 10K+ requests/day at <200ms p95 latency
- Deployed ML microservices: spaCy resume parser (95% accuracy) and real-time CV pipeline at 30 FPS
| Project | Description | Stack |
|---|---|---|
| 🌐 P2P Event Mesh | Decentralized event mesh with Chord DHT — zero SPOF, tested on Amazon EC2 | Java · gRPC · Protobuf |
| ⚡ Pulse Orchestrator | Distributed task orchestrator — 10K+ tasks/day, 99.9% reliability | Python · FastAPI · Redis |
| 📈 Dummy Trading | Paper trading with ML ensemble predictions (LSTM + GRU + Holt-Winters) | TypeScript · Node.js · TF.js |
| 🚗 Vehicle Matching | CV pipeline — 94% accuracy across 100K+ images, GPU-accelerated | Python · YOLO · CUDA |
| 🤖 Credit Risk System | XGBoost inference — 91% AUC-ROC, real-time Flask API <200ms | Python · XGBoost · Flask |
| ☁️ CF Doc Explorer | Semantic search on Cloudflare edge — 70ms cold start, BGE + Gemini | TypeScript · CF Workers |
Always building. Always shipping.