Let'sConnect
Software engineer and ML researcher, based in Dhaka.
Final-Year CSE Student at UIU | Full-Stack Developer | ML Researcher
I build production software and do applied machine learning research. I'm finishing my B.Sc. at United International University with a 3.94 CGPA and eight merit scholarships so far, six of them at 100%, and I work there as a teaching assistant. My FYDP-II thesis trains a physics-informed network on the ODE residual alone, with side studies in physics-based attack detection and AI-image forensics.
What I'm doing nowFeatured Projects
View All ProjectsE-Learning Platform
The live IELTS preparation platform I engineer solo for RimonsIELTS: a Next.js 15 student portal with mock tests and video courses, an Express + TypeScript API with enforced OpenAPI contract parity, a React admin with a full test pyramid, and a Python central service. Live at rimonsielts.com.
Advyon
Legal-tech platform with case management, document workflows, and AI-assisted legal insights, built on a feature-module React 19 architecture. I led a five-member team through the full SDLC and we won Champion at the UIU CSE Project Show.
Lorenz ODE Thesis
My FYDP-II thesis, defended: a physics-informed network that never sees the answer, trained on the Lorenz-1960 ODE residual alone, lands within 3e-05 of a DOP853 reference in under six CPU minutes. Nine architectures swept, sixteen paper runs archived, 41 tests, manuscript submitted to ICEEICT 2027.
Skills Overview
I work across web development, backend systems, and applied machine learning. Most of it was built through shipped client work, university team projects, and research.
Languages
Frontend
Backend
ML & Tools
From the Blog
View All Posts- 01Rules You Can Read: An Adaptive Rule-Based Classifier for Biological DataBiological data often means many features and few examples, along with noise and contradictions. Farid et al.'s 2016 classifier combines random feature subsets, error-focused reweighting, and C4.5 rules so the result stays readable, then checks its leftover mistakes with weighted kNN.AI and Machine LearningOctober 8, 2026 · 7 min read
- 02Two Hybrids, One Idea: Pairing Decision Trees with Naive BayesFarid et al.'s 2014 paper builds two independent hybrids. One lets naive Bayes delete the training rows a tree would otherwise overfit. The other lets a tree select and depth-weight the attributes naive Bayes uses. Same partnership, run in opposite directions, with one caution about what a model's mistakes really mean.AI and Machine LearningOctober 5, 2026 · 7 min read
- 03Reading a Moving Target: The Adaptive Ensemble for Concept-Drifting Data StreamsA dataset is a snapshot, but a stream keeps arriving and its meaning shifts over time. Farid et al.'s 2013 adaptive ensemble holds three weighted decision trees, focuses later trees on the cases earlier ones missed, watches its leaves for a new class, and swaps the weakest tree for a fresher one. This is how the pieces fit.AI and Machine LearningOctober 2, 2026 · 7 min read
Now · November 2026
The present tense
What I'm building, teaching, and researching right now, in the short version. The /now page always carries the full picture.
Schedule a Meeting
Book a 15-minute call to talk about a role, a project, or a possible collaboration. Message me on WhatsApp and we'll find a time.
Schedule NowNewsletter
Occasional updates on new projects, writing, and research.