$ whoami
Senior Software Engineer at Enosis Solutions · B.Sc. in CSE, AUST
I build reliable, data-driven software. For the past 3+ years I have designed and shipped full-stack enterprise systems (project management, payment approval and workspace reservation platforms) with TypeScript, NestJS, React and PostgreSQL.
My research goal is to make machine learning reliable where data is scarce: models that learn well from limited or simulated data, are tested honestly against simpler baselines, and turn complex scientific data into insights people can act on. This grew out of my IEEE paper on Bengali–English neural machine translation, and now drives my independent research in photonic sensor modelling and eDNA-based biodiversity monitoring.
Seeking graduate research opportunitiesLearning from limited or simulated data with generative augmentation, evaluated rigorously against simpler baselines.
Machine translation and Bengali NLP with deep learning where parallel data is scarce.
Data-driven modelling of physical sensors and turning ecological data into interpretable insights.
GAN-Augmented Neural Network Modelling of a New Open-Channel PCF–SPR Plasmonic Sensor
Aims to compare ANN and XGBoost models, each with and without WGAN-GP data augmentation, for predicting the performance of a new open-channel photonic crystal fibre surface plasmon resonance sensor.
From eDNA to Insights: An AI-Driven Environmental Monitoring Platform for Biodiversity Analysis
Aims to develop an AI-driven platform that turns environmental DNA (eDNA) survey data into interpretable biodiversity insights for environmental monitoring.
Bengali-English Neural Machine Translation Using Deep Learning Techniques
2023 3rd International Conference on Electrical, Computer and Communication Engineering (ECCE), CUET, Chittagong, Bangladesh. IEEE, 2023.
@inproceedings{paul2023bengali,
title = {Bengali-English Neural Machine Translation Using Deep Learning Techniques},
author = {Paul, Nipun and Faruki, Ishmam and Pranto, Mutakabbirul Islam
and Shawon, Md. Tanvir Rouf and Mandal, Nibir Chandra},
booktitle = {2023 3rd International Conference on Electrical, Computer and
Communication Engineering (ECCE)},
year = {2023},
doi = {10.1109/ECCE57851.2023.10101491}
}Decentralized Election System: Blockchain and Crypto Based Vote Transaction
B.Sc. Thesis, Department of CSE, Ahsanullah University of Science and Technology.
A secure, transparent and tamper-proof voting system on Ethereum, designed for the context of Bangladesh, that uses a distributed ledger to provide election integrity, voter anonymity and real-time verifiability.
Enosis Solutions
Project Management System · NestJS, React, PostgreSQL
Payment Approval System · PERN stack
Desk Reservation System · PERN stack
Projects I am building now.
Strava-style GPS activity tracker; segment efforts are matched with spatial SQL in PostGIS.
Seat-reservation platform with zoomable venue maps, a “view from your seat” preview and analytics.
Ahsanullah University of Science and Technology (AUST), Dhaka
CGPA 3.38 / 4.00 · 3.66 over the final five semesters · Thesis: Decentralized Election System (Blockchain and Crypto Based Vote Transaction)
Rajuk Uttara Model College, Dhaka
GPA 5.00 / 5.00
Anjuman Adarsho Govt. High School, Netrokona
GPA 5.00 / 5.00
Bengali: native · English: IELTS Academic 7.0 (Listening 7.5, Reading 8.0, Writing 6.5, Speaking 6.0)
I’m happy to talk about research collaborations, graduate opportunities or engineering work. Email is the fastest way to reach me: