Skip to content

Daksh Saijwal

Mechanical Engineering @ IIT Kanpur

Most of my time goes into building software: a multiplayer typing game, an air quality forecasting model for Delhi, and a linkage design tool for an ME course. Codeforces Expert on the side.

GitHubContact
▼ begin voyage

01/About

About me

I'm a mechanical engineering undergrad at IIT Kanpur. On paper, at least. Most of my actual hours go into writing software.

Recent builds: a multiplayer typing game with live races, a model that forecasts Delhi's air quality, and a tool from one of my ME courses that designs four-bar linkages in the browser. There's no grand theme. I build whatever looks interesting and see it through.

Outside of that I do competitive programming on Codeforces (Expert, 1608) and watch a lot of Formula 1.

Studying
Mechanical Engineering, IIT Kanpur
Working on
Full-stack platform at Apni Vidya
Comfortable in
C++, Python, JavaScript / React
For fun
Competitive programming (Codeforces Expert)

02/Journey

The route so far

The jumps that got me here.

> route plotted · 10 jumps · 12 parsecs

> leaving charted space · thanks for flying along

▼ scroll to fly

03/Skills

What I work with

Only the stuff I've actually used in projects, nothing padded.

C++· Competitive programming
Python
JavaScript

04/Work

Things I've built

Click any of these for the full write-up.

Real-time Systems2025 · Active

A typing race game with live multiplayer over Socket.io, ghost replays of your personal best, and per-key analytics that surface the characters slowing you down.

Realtime
Socket.io
Difficulty modes
3
Analytics window
20 races
WPM chart
2 Hz
ReactViteJavaScriptSocket.ioNode.jsExpressTailwind CSSMongoDB
Machine Learning2025 · Shipped

A leakage-safe ML pipeline that forecasts hourly concentrations of six pollutants across Delhi and renders 7-day, city-wide AQI heatmaps using gradient-boosted models. Won the IIT Kanpur research hackathon.

PythonLightGBMXGBoostPandasNumPy
Computational Design2025 · Shipped

A framework-free web tool that solves the inverse kinematics problem: given the points a coupler must pass through, it synthesizes physically valid four-bar linkages that trace them. Built for ME252.

JavaScriptHTML5 CanvasWeb WorkersCSS
More on GitHub

05/Experience

Experience

Internships so far. Expand a role for details.

Building a multi-role web platform that automates how coaching institutes run: enrollment, attendance, fees, testing, and reporting across admin, teacher, student, and parent roles.

  • Built a Node.js/Express REST API (16 modules, 80+ endpoints) on PostgreSQL with JWT auth and route-level role-based access across four roles.
  • Designed a QR-driven enrollment pipeline that auto-provisions linked student and parent accounts with one-time passwords in a single transaction.
  • Built an MCQ engine with server-side auto-grading, configurable negative marking, live re-ranking, and per-topic gap analysis feeding student SWOT reports.
  • Integrated Razorpay payments with HMAC-SHA256 verification and idempotent, race-safe reconciliation (SELECT FOR UPDATE) so webhooks and client-verify cannot double-credit.
  • Built a pluggable notification layer (MSG91 SMS, Gupshup WhatsApp) with full dispatch audit logging and a scheduled daily/weekly/monthly report dispatcher.
  • Hardened auth: bcrypt-hashed passwords and OTPs, anti-enumeration responses, constant-time HMAC compares, and an env validator that refuses unsafe production config.
ReactNode.jsExpressPostgreSQLJWTRazorpay

06/Contact

Get in touch

The form goes straight to my inbox. Or just email me directly.

dakshsaijwal06@gmail.com