I’m Harsh N. Chandak,
• MSCS student at ASU, Tempe
• Experienced Full-stack Developer
• Software Development
• Database Management
(My name means "Happiness" in Hindi. Yes, I'm aware of the irony.)
"Good code is like a good joke: if you have to explain it, you messed up."– Harsh N. Chandak
// About Meconst dev = {
name: "Harsh N. Chandak",
title: "Full-stack | Backend Developer",
education: "MSCS @ASU",
philosophy: "Ship if chaotic, automate if boring, scale before needed",
core_skills: [
"REST APIs",
"Web Dev",
"Data Viz",
"System Optmization",
"System Design"
],
traits: [
"system-thinker",
"api-gardener",
"legacy-fixer"],
daily_setup: [
"☕ Coffee",
"Postman",
"Jira",
"VS Code",
"browser tabs",
"notes"],
tech_stack: [
"JavaScript",
"TypeScript",
"Node.js",
"React",
"Next.js",
"Express.js",
"PostgreSQL",
"MongoDB",
"MySQL",
"Python",
"HTML",
"CSS",
"Redis",
"Git",
"Docker",
"AWS"]};
// A personal project born from job hunt frustration.
Used Puppeteer + cron + MongoDB with JWT-auth to personalize tracking and statuses. Saving hours of weekly job hunt efforts for real-time discovery.
// ASU course project on distributed systems
Built a Kafka → Neo4j stream processor with Dockerized BFS/PageRank analytics. Enabled live user graph insights with real-time updates under load.
// ASU Knowledge Representation and Reasoning project
Simulated warehouse robot logic using Clingo + Python wrapper. Solved multi-agent delivery plans with optimal horizon search in ASP.
// ASU big data and geospatial analytics course
Analyzed large spatial datasets via Spark + Scala + Spatial SQL. Optimized joins, geohash clusters, and RDDs for fast geospatial queries.
// ASU D3.js data storytelling project
Mapped 185K+ accident records using D3.js & GeoJSON. Built six interactive charts, placing 3rd in class data storytelling showcase.
// ASU research project exploring AI and blockchain synergy in cybersecurity
Studied AI + Blockchain for breach detection in fintech. Proposed verifiable audit trails using anomaly detection + smart contracts.
// Undergraduate academic project focused on secure property ownership management.
Built a DApp for property ownership using Solidity + Flutter. Tested secure record transfers on Ethereum (Ganache + Truffle).
// Undergraduate academic project
Automated OMR grading with OpenCV under noisy lighting. Tuned bubble detection + GUI preview for accuracy and flexibility.
// Undergraduate capstone project
Trained a FFNN to classify sign gestures using MediaPipe features. Achieved ~98% accuracy in real-time hand-to-text prototype.
Taming Complexity and operations, from sales to dispatch.
️ Built inventory logic that auto-rebalanced stock across warehouses, yes, it had opinions.
Designed BI dashboards that told stories, not just numbers (hello, exec summaries).
️ Production planning? Gantt-inspired timelines that respected both dev cycles and machine breakdowns.
Assumed failure as the default, built like it.
If you gave me your ops chaos today, I’d model it by Monday and stress test it by Friday.
It isn't “build a store.” It is: make commerce sing.
Shipped 15+ full-stack storefronts. From fashion, accessories, to hardware and spare parts.
Integrated payments (Razorpay/Stripe) with grace, retries, refunds, and regrets included.
Backend: custom schemas per vertical. Frontend: fast enough to feel native.
Dashboards that helped sellers sell, not confuse them.
Juggled orders, stock, delivery, roles, and edge cases without dropping the vibe.
Real-time isn't speed, it's trust. And I build like I mean it.
Location tracking with socket + fallback polling, accurate and battery-friendly.
Chat apps with sticky state, retry queues, and graceful reconnects.
Built for poor networks. Always. Because people use phones, not fiber.
UI that reacts fast but never freaks out. Real-time should feel smooth, not sweaty.
Debugged packet drops with browser tabs and sheer will.
Learning platforms are a mess. I made them less so.
Built progress trackers that knew when to chill (don’t panic over skipped quizzes).
Admin tools that didn’t require a handbook, drag-drop, done.
️ Real-time session tracking without WebSocket meltdowns.
Modular content pipelines, one backend to teach them all.
Asked myself: “Would a teacher actually use this?”, and designed from that answer.