PhD – Machine Learning – 2022-2025

From – Federation University Australia (Soil CRC Scholarship)

Thesis: Enhancing Soil Organic Carbon Estimation through Hyperspectral Dimensionality Reduction.

Examiner feedback:

“This is an outstanding, and highly complete piece of doctoral research.”

“… goes beyond black-box performance comparison, offering deep insights into the algorithms’ inner workings.”

“… and the academic style is impeccable.”

BSc – Computer Science – 2004-2007

From – American International University-Bangladesh

Awards: Summa Cum Laude and scholarship for academic excellence.

Analyst Programmer

06/2025 – till date

From – Wimmera CMA, Australia

Leading GIS and software development initiatives, delivering data-driven environmental and geospatial solutions using Python, remote sensing, and machine learning.

Machine Learning Engineer

08/2022 – 05/2025

From – WebAlive, Australia

Developed machine learning–driven web applications, including semantic search systems and computer vision pipelines for real-world business workflows.

Engineering Director

07/2008 – 04/2022

From – Bit Mascot, Bangladesh

Led and architected full-stack development across diverse projects, from customised deployment platforms to enterprise billing systems.

Interactive GIS application development

Features: Cross-platform apps delivering interactive, user-friendly GIS layers

Skills: QGIS Server management, GIS data editing, GIS protocols, Bash scripting, mobile app development

Tools/Platforms: Ubuntu, Azure, Cursor IDE, Android Studio, Xcode, QGIS

Spectral index analysis and comparison system

Features: Interactive visualisation of spectral indices (NDVI, NDWI, etc.) across timelines with spatial point-based comparison

Skills: Remote sensing analysis, time-series processing, geospatial data handling, interactive visualisation

Tools/Platforms: Python, Google Earth Engine, React, AWS

ML system for vegetation classification

Features: Classification of native vs commercial vegetation

Skills: PyTorch model development, first-principles computer vision, pipeline design, data annotation strategy, training and evaluation protocols

Tools/Platforms: PyTorch, web-based interfaces for model testing

GIS reporting & spatial analytics

Features: Decision-ready spatial reports with precise filtering and validation

Skills: Spatial queries, spatial joins, overlay analysis, geoprocessing workflows, requirement interpretation, accuracy validation, automation with AI agents

Tools/Platforms: ArcGIS Pro, QGIS, Python, SQL, Claude

IEEE Access (2025)

Deep Learning-based Adaptive Downsampling of Hyperspectral Bands for Soil Organic Carbon Estimation

Sensors 24 (23), 7771

BSDR: A Data-Efficient Deep Learning-Based Hyperspectral Band Selection Algorithm Using Discrete Relaxation

How I Am