[ CURRENT STATUS ]
KUET · ECE · SHIFTING DEEP INTO ML & DATA SCIENCE
[ PERSPECTIVE ]
"learning by breaking things, then rebuilding them better."
I am an undergraduate student of Electrical and Communication Engineering at KUET, exploring at the intersection of hardware constraints and predictive software modeling. My approach is fundamentally vertical: I believe in understanding the math behind neural networks as deeply as the physical parameters of signal-filtering sensors.
I've built a variety of systems — a syllabus tracking platform, an Arduino-based desk utility, and active-reading study tools — primarily to push my limits and see what breaks along the way. These days, most of my focus is directed toward Machine Learning and Data Science as a natural next chapter.
[ SELECT NODE OR SWIPE / SCROLL CARD ]
Internships at Siliconx Softwares and Firefly Hi-Tech. Provided a firsthand look at how production codebase architectures and hardware systems are maintained under team constraints.
[ SCROLL TO SLIDE CARDS ]
An early dip into industry software development — focusing on understanding production web architectures, team code environments, and modern development workflows.
Hands-on hardware and web work, built HTTP IoT monitoring dashboards, fixed live payment gateway issues, and assembled Arduino hardware prototypes for real-world tests.
Joined as co-founder and CTO of Protiva — a local-first, AI-native PDF workspace engineered around active reading, context-aware explainers, and live study rooms for students.
[ SELECTED SHIPS & PROJECTS ]
[ SKETCHBOOK PREVIEW ]
"Not every project needs a stage. Some were just reps."
A list of 10+ command-line tools, Progressive Web Apps, e-commerce experiments, and our software graveyard.
[ COMPETITION WINS, RECOGNITIONS & VERIFIED RECEIPTS ]
"Pitched a mobile waste-to-green-coal processing system designed to eliminate factory waste transport on-site."
Presented "Oratory" concept at Hult Prize's KU round, addressing massive logistics inefficiencies in industrial waste transport.
"Wrestled with messy, high-cardinality dropout risk metrics to optimize predictive performance."
Competed in Byte Datathon by K-MINDS, handling missing datasets, feature engineering, and fine-tuning log loss.
Lately, my core focus has shifted toward Machine Learning and Data Science. I am actively exploring predictive models, gradient descent optimization, and neural network architectures — linking mathematical foundations back to embedded hardware and full-stack web systems.
Got an idea, a project to build, or just want to chat about machine learning and IoT? Send a signal.