I'm 22 and studying Information Technology at Guru Nanak Institute of Technology in Kolkata. Most of what I make is small — command-line tools, little services, bits of glue that save me repeating myself.
Nearly everything here started as something I wanted for myself. gitto exists because I kept forgetting git commands. Bhoomi started because serving satellite imagery looked hard and I wanted to find out whether I could. That has been a better reason to start than trying to pick something impressive.
I usually have Claude and Cursor open these days, so less of my time goes on typing and more on working out whether what came back is any good. I'm not always right about that.
I have probably learned more from reading other people’s code than from any course I have taken. Trying to work out why someone made a particular choice — and whether I would have thought of it — sticks with me longer than being handed the answer.
What I care about is whether a thing holds up once someone else is using it. Working on my laptop and working when two people click at the same time turn out to be quite different, and I keep learning that the slow way.
Outside coursework: a couple of small packages on npm, a service running on Render, an NPTEL Java certification I am quietly pleased about, and a paper on recommender systems I co-authored in June 2026.
Mostly though, I make things and find out what breaks. That has taught me more than anything else so far, if not always gently.
What I use
Mostly Python and TypeScript. FastAPI or Node.js when something needs to keep running, PostgreSQL when the data has a shape to it, Redis when something is too slow to do inside a request. Next.js and Tailwind CSS when a project needs a face. Docker and Linux because everything ends up there eventually, and pytest because I have been burned. I'm not equally comfortable in all of these — some I've used just enough to get something working.
Record
Work
Association for Computing Machinery
Student member
ACM
IIP Series
Co-authored a chapter
Advances in Machine Learning, Volume 6S1 — peer reviewed
- Adaptive Cluster-Smoothed SVD: A Hybrid Recommender System with Dynamic Weight Optimization.
- Matrix factorisation plus dimensionality reduction, aimed at rating prediction when the user-item matrix is mostly empty — which it usually is.
Courses and certifications
Things I sat down and finished
- Programming in Java — NPTEL, IIT Kharagpur. Elite and Top 5%, consolidated score 99%.
- Data Analytics with Python — Institute for Advanced Learning and Skill Development, March to April 2026.
- Web Development with Node.js and AngularJS — Institute for Advanced Learning and Skill Development, July 2026.
Education
Guru Nanak Institute of Technology
B.Tech, Information Technology
Kolkata — affiliated to MAKAUT, West Bengal
- CGPA 8.42 / 10 so far.
Barasat I. G. Memorial High School
CBSE Class XII
Physics, Chemistry, Maths, Computer Science
- 75.2% in Class XII, 90.0% in Class X.
Kolkata, India — roynayanroy14@gmail.com