Software engineer & AI enthusiast
Based in ChattogramBangladesh · UTC +06
I’m a software engineer focused on Web Experimentation and CRO Development, with experience building dashboards and automating workflows.
I’m an AI enthusiast interested in model selection: comparing accuracy and error costs, especially when classes are imbalanced.
I studied computer science at CUET. My undergraduate work on landslide susceptibility led to a paper on relative model-selection metrics.
Visits appear here once the site is live
Home base in dark red
City-level estimates · About this map
Applying for CS PhD programmes for Fall 2027, with broad interests in Artificial Intelligence, Machine Learning, and Software Systems.
Web Experimentation
Follow visitors through two checkout designs. Watch completed orders collect, and see the results take shape.
This illustrative experiment sends 100 visitors to each checkout. A produces 52 completed orders and B produces 68. Enable JavaScript to watch visitors move through the pages and build the 3D result columns.
Research & engineering
Rangamati
Hill Tracts, BD
Undergraduate thesis · CUET
A Flask application that compares seven models using eight environmental factors, assesses coordinates within Rangamati, and maps susceptibility across the district.
Python / scikit-learn / Flask / Leaflet
Publication · IEEE RAAICON, 2021
A scoring rule that combines standardized accuracy and error across a candidate pool. The research notes explain the formula and how adding or removing a candidate can change the ranking.
Engineering · Client project
A React dashboard showing experiment release frequency, development hours, and team progress from Jira data. Built to help teams monitor their workflow and plan sprints.
React / Node.js / Express / Jira REST API
Production tooling / Data flow
Webhook → Validate origin → Sync update
Idempotent writes. No echo loops.
Engineering · Client project
A Python webhook service that keeps tasks, statuses, and timelines in sync between Airtable and Jira, reducing manual re-entry in experiment tracking.
Python / Airtable API / Jira REST API / Webhooks
Research interests
My thesis raised a question I’m still interested in: how should we compare models when false positives and false negatives have different consequences?
How should a selection rule express the different consequences of false positives and false negatives?
What should we measure when the rare event matters most, and aggregate accuracy hides it?
How do we choose between models when correctness, latency, and cost are separate objectives?
I’m based in Chattogram, Bangladesh. At EchoLogyx, I work on Web Experimentation and CRO Development and have designed and analysed around 800 web experiments on production traffic.
My engineering projects include a workflow metrics dashboard and a two-way Airtable–Jira integration for experiment tracking.
Earlier interests included robotics and mathematics: a line-following robot competition win and a regional Mathematical Olympiad podium.
For research discussions, engineering work, or PhD opportunities: