SyedNazmusSakib

I build things that work, then keep poking at them until they work better. I'm a software engineer and ML researcher studying at EDU, part builder and part researcher. I read the papers and ship the code. I'm a fast learner and a stubborn problem-solver, happiest with a messy problem I can't quite put down yet.

Most of my research time goes into one question: where are Bangladesh's fast-growing cities losing their green spaces, and can we predict it before it happens? I work on urban vegetation prediction, temperature increase and green forest ecological balance using satellite embeddings and ensemble ML. My approach is hands-on. I read and review research papers, pull out the information and findings that matter, then jump straight into experimenting with models and fine-tuning their predictions. I care about systems that are accurate and explainable, because a prediction you can't understand isn't worth much.

When I'm not researching, I'm building full-stack web apps that solve real problems. EDU ClassRepo, TurfCTG, Foshol Bachao, and Taj, each one tackling something practical. Building them end-to-end keeps me sharp across the whole stack and reminds me that good engineering is really about people, not just code.

I stay sharp through competitive programming and a genuine love for the hard stuff. A new ML technique, a tricky system to design, a brutal algorithmic problem. I dig in with curiosity and stubbornness in equal measure. Next up: a master's abroad, to take this research further alongside experienced people I can learn a lot from.

Outside all that, I'm drawn to unfamiliar streets and a good chess endgame. You'll catch me on the field with a football or cricket bat, swinging a badminton racket, or thinking three moves ahead over a board. It keeps me balanced, a little competitive, and always learning. Not done yet, just getting a little better every day.