I solve problems end to end: the code, the marketing, the sales calls,
whatever the day actually needs. Currently heads-down on
AtlasLinq, an NFC business-card platform,
still running in production.
const nuzair = {
title: "Software Engineer, technically",
building: "AtlasLinq — product, code, and everything in between",
base: "Indore, India",
daily: ["too much coffee", "the gym", "a build pipeline that owes me an apology"],
};
There's a terminal hiding on this site. Press ~ if you're
the type who checks. Or use the switch in the corner.
work
AtlasLinq (Oct 2024 – present)
NFC digital business cards. Founded, designed, and built alone: mobile app,
web dashboard, and a Firebase backend running 35 deployed cloud functions.
The "tap" part is a real native Android NFC Host Card Emulation service, not a
third-party SDK doing the work for me. Real payments, multiple providers, live
in production. I verified the checkout flow actually charges real money by
charging myself ₹249. It worked. Mildly anticlimactic, which is exactly what
you want from a payment flow.
Also shipped: a script that fails my own build if the web and
backend permission matrices (RBAC) drift out of sync. It has caught real bugs.
It has also ruined a few otherwise fine afternoons.
AtlasLinq — the OCR pipeline
A 7-stage, confidence-scored fusion pipeline for parsing business cards and
IDs offline. It handles Indian and international name/address formats, which
turned out to be its own small research project. Replaced an earlier
version that once flagged someone named "Chris" as a job title, because a
naive keyword check matched the letters "h" and "r" sitting right inside his
own name.
GloFi (fintech contract, May 2025 – Feb 2026)
Owned the client-facing dashboard and website end to end, including real
GA4 analytics, JWT auth, and CSV/PDF export. Also shipped a Transak fiat
on-ramp integration, a gamification system, and a contacts/QR feature on the
wallet app itself.
The bigger lesson wasn't technical. A good idea with a
scattered team gets you a good demo, not a good product. No amount of clean
code fixes a leadership problem. I weigh "who's actually in the room and
pulling together" as hard as the idea itself now, before I say yes to anything.
Sustainable Shopper (Jun 2025)
gpt-4o (multimodal) + CLIP/FAISS embedding search, for matching outfits to
more sustainable alternatives. The vision model reads a photo plus live
weather data and suggests something you'd actually wear that day. Virtual
try-on is outsourced to Fashn.ai, since there's no reason to pretend I solved
that part too.
Kratos (Apr 2026 – present)
Built primarily for my mom, after she started going to the gym and wanted a
straight answer to "what am I actually supposed to do today." Kratos gives her
a real, predefined workout and walks her through it. It even syncs to a
home-screen widget, so she can check today's plan without opening the app.
The on-device pose-detection layer (Flutter + Riverpod + ML Kit) came in for
me and my brother: a heat-map view of what's actually getting worked, the part
that turns a to-do list into a coach.
When mom says go work out, you build the app that makes sure
she works out. No follow-up questions asked.
Philips — Synergy GUI
Software development engineering internship on hospital patient-monitor
software: TypeScript, Lit web components, and Philips' internal DLS4 design
system. Independently verified: my mentor's written review rated me top 1–5%
overall, 10/10 on motivation and communication. Not my own claim, that one's
on record.
No hospital wing named after me yet, but a monitor I helped
design the UI for is out there right now, quietly keeping an eye on someone,
somewhere. I'll take it.
about
Software engineer, technically. I think of myself as a problem solver first.
Founder of AtlasLinq, based in Indore, India. I like building the whole thing,
front to back: the product, the code, the marketing, the sales calls, all of it.
Staying in one lane was never really the plan.
Played a lot of team sports growing up, football mostly, seriously enough
that some of it stuck. I still default to treating problems like something you
iterate on with a team and a few bad first attempts, not something you solve
alone in one clean shot.
Outside of work: gym, gaming, music, reading, some photography.
No further comment.
Daily tools: Claude Code and Antigravity, for research, refactors, and the
parts of engineering nobody puts on a highlight reel. Currently also building
Braces Reborn, an MCP server, on the
side.
sidequest scratchpad
Half build-log, half "things I got briefly obsessed with."
Updated whenever, not on a schedule.
currently
Wiring an MCP server (Braces Reborn) into an actual agent loop instead of
calling it by hand like an animal.
side quest
The RBAC parity-check script mentioned under AtlasLinq started as a
20-minute "let me just check something." It is now a permanent part of the
build. This happens more often than I'd like to admit.
reading
stub — fill in later. "Later" is doing a lot of work in that
sentence.
blog
buffering...
(Still is, actually. Check back later.)