Kunj Pandya

Full-Stack Engineer & Entrepreneur

📍 Hoboken, NJ

Building scalable backend systems and intelligent applications. From distributed task orchestrators to AI-powered platforms, I turn ideas into production-ready solutions that scale.

10K+
Tasks/Day Processed
500+
Active Users
94%
ML Accuracy
View Resume

About Me

Full-stack engineer and aspiring entrepreneur passionate about building scalable systems that solve real-world problems.

Currently pursuing MS in Computer Science at Stevens Institute of Technology while architecting distributed backend systems, microservices, and cloud-native applications.

From building distributed task orchestrators handling 10K+ tasks/day to creating AI-powered platforms serving 500+ users, I focus on turning ideas into production-ready solutions that scale.

Education

Master of Science

Computer Science

Stevens Institute of Technology

Hoboken, NJ

Sept 2025 – May 2027

Key Coursework:

Distributed SystemsCloud ComputingConcurrent Programming+2 more

Bachelor of Technology

Artificial Intelligence and Machine Learning

D. J. Sanghvi College of Engineering, Mumbai University

Mumbai, India

Sept 2021 – May 2025

Key Coursework:

Machine LearningDeep LearningComputer Vision+2 more

Experience

Backend Engineer – Distributed ML Platform

Current
Stevens Institute of Technology
Hoboken, NJ

Nov 2025 – Present

  • Architecting and building a distributed profiling platform for federated machine learning, handling energy metrics collection from 100+ edge devices in real-time.
  • Developed scalable backend infrastructure using TensorFlow.js, Socket.IO, and Redis to process per-layer power consumption data across heterogeneous device networks, optimizing for low-latency data aggregation and fault tolerance.
TensorFlow.jsSocket.IORedisDistributed SystemsReal-time ProcessingEdge ComputingSystem Architecture

Founding Engineer – Placement Intelligence Platform

D. J. Sanghvi College of Engineering
Mumbai, India

May 2024 – June 2025

  • Built and scaled backend infrastructure from scratch for a placement platform serving 500+ students. Designed RESTful APIs handling 10K+ requests/day with <200ms p95 latency, optimizing database queries and implementing efficient caching strategies.
  • Architected and deployed ML inference microservices including an spaCy-based resume parser (95% accuracy) and real-time computer vision pipeline processing at 30 FPS. Managed end-to-end product development from initial prototype to production deployment.
Node.jsSQLiteRESTful APIsMicroservicesspaCyTensorFlowSystem DesignProduct Development

Featured Projects

A selection of projects demonstrating expertise in distributed systems, AI/ML, and scalable backend architecture

Full-Stack & Machine Learning

Dummy Trading — Paper Trading Platform

Feb 2026

Comprehensive paper trading platform with $100K virtual funds, real-time Alpaca pricing, ML-powered predictions, and competitive group trading.

Impact: Full S&P 500 market browsing with real-time pricing. ML prediction system with 23-feature engineering pipeline from OHLCV data. Supports limit/stop orders with automated 60-second validation cycles.

ReactTypeScriptNode.jsPostgreSQLPrisma+3 more
Game Development & AI

Wizard Duel — 3D Battle Game

Feb 2026

First-person 3D wizard battle game built with Unity 6 and URP, featuring a 10-level Harry Potter-themed campaign with voice-controlled spell casting.

Impact: 10-level campaign from Professor Quirrell to Lord Voldemort across themed arenas. State machine AI with weighted decision-making and dynamic arena theming per level.

Unity 6C#URPCinemachineWhisper API+2 more
Computer Vision & Deep Learning

Vehicle Matching System (AIRGarage Track)

Dec 2025

Multi-stage computer vision pipeline for associating vehicle entry-exit events across 100K+ images with 94% matching accuracy.

Impact: Achieved 94% matching accuracy across 100K+ images. Reduced end-to-end processing latency under batch workloads using GPU acceleration.

PythonYOLOFastALPROCRResNet+3 more
AI & Edge Computing

Cloudflare AI Doc Explorer

Jan 2026

Production-ready semantic documentation search and AI chat system built on Cloudflare's edge infrastructure with vector embeddings and Gemini.

Impact: 70ms cold start, 200-400ms search latency. Bundle optimized from 1.5MB to 285KB gzipped. 25+ indexed documentation chunks with persistent chat history.

TypeScriptCloudflare WorkersVectorizeGeminiReact 19+2 more
Machine Learning & Backend

Credit Risk Inference System

Jan 2026

Production-grade credit risk model using XGBoost analyzing 50,000+ borrower records with 91% AUC-ROC and real-time Flask inference API.

Impact: Achieved 91% AUC-ROC and 81% accuracy. API responses under 200ms median latency, handling 1,000+ daily predictions with concurrent request support.

PythonXGBoostscikit-learnFlaskPostgreSQL+2 more
Security & Backend

PassManager Server

Jan 2026

Lightweight password manager server with trust-free architecture, end-to-end encryption, and secure session management.

Impact: Lightweight deployment suitable for small-scale infrastructure. Complete cryptographic implementation with secure authentication and encrypted data storage.

PythonCryptographyREST APISession ManagementDatabase
AI & Full-Stack

VentureViewHR — AI HR Assistant

Nov 2025

AI-powered HR analytics application built with Gemini AI for intelligent recruitment insights and workforce management.

Impact: AI-driven HR analytics with real-time Gemini-powered insights for recruitment and workforce management decisions.

TypeScriptGemini AIReactNode.js
Distributed Systems

Pulse — Distributed Task Orchestrator

Oct 2025

Horizontally scalable distributed task orchestration service with REST APIs, Redis priority queues, and fault-tolerant task processing.

Impact: Handles 10K+ tasks/day with 99.9% reliability. Resolved task starvation and retry storms under burst workloads, stabilizing end-to-end task processing.

PythonFastAPIPostgreSQLRedisDocker+1 more
Distributed Systems & Networking

Decentralized P2P Event Mesh

Aug 2025

Decentralized peer-to-peer event mesh using Java, gRPC, and Chord-based DHT for real-time push-based message propagation.

Impact: Evaluated on Amazon EC2, maintaining consistent message delivery during node joins, failures, and migrations. Zero central point of failure with automatic DHT rebalancing.

JavagRPCEC2ProtobufChord DHT+2 more
AI/ML & Backend Systems

SmartKitchen AI

Oct 2024

Scalable backend with MobileNetV2-based ingredient recognition, recipe catalog of 8,000+ entries, and hybrid recommendation engine.

Impact: Real-time ingredient recognition and personalized recommendations. Automated inventory tracking and nutritional data retrieval in end-to-end production workflow.

PythonMobileNetV2TensorFlowTF-IDFBackend Systems+1 more

Technical Skills

Languages

Java (Proficient; Concurrency)Python (FastAPI, NumPy)C/C++ (CUDA, Systems Programming)JavaScript/TypeScriptGo

Backend Systems

Distributed Systems (DHTs, Fault Tolerance)gRPCRESTful APIsPostgreSQL (Schema Design)RedisMicroservices

Infrastructure & ML Ops

DockerAWS (EC2/S3)Model Inference PipelinesLinux/BashCI/CD PipelinesGit

AI/ML

TensorFlowPyTorchComputer Vision (YOLO, ResNet)NLP (spaCy, NER)Deep LearningRecommendation Systems

Developer Tools

GitUnit Testing (JUnit, PyTest)Maven/GradleVS CodeDebugging & Profiling

Always learning and exploring new technologies to build better systems

Get in Touch

I'm always open to discussing new opportunities, collaborations, or just having a chat about distributed systems and AI/ML.

© 2026 Kunj Pandya. Built with Next.js, TypeScript, and Tailwind CSS.