>_ specializing in
I'm an AI Engineer & Systems Architect specializing in computer vision, geospatial data engineering, and photogrammetry. Over the last three years at HawarIT Limited, I've designed and shipped production AI systems that take raw aerial and satellite imagery through cleaning, correction, and ML inference to products people actually make decisions with — orthophotos, DEMs, quality-control reports, crop and change maps. In industry, the model is the easy part; the data pipelines around it are the system. That's the part I architect.
ml_systems/
End-to-end ML systems — from benchmarking and training models to serving them as independent microservices behind REST APIs.
data_eng/
The pipelines where production ML is won: preprocessing, postprocessing, ETL, and QC across terabyte-scale imagery.
architecture/
Designing the whole system — modular libraries, distributed services, edge deployments, and tool interoperability.
When I'm not coding, I enjoy playing video games, watching movies and spending time with my family and friends. I also love playing football, it has been a big part of my life in my teenage years. I enjoy learning new things. I am currently learning about islamic history and philosophy. I used to love working out and am getting back into it as well.
Always looking forward to new opportunities to learn and grow. Let's collaborate and bring your next project to life!
Desktop application that decodes drone thermal imagery into per-pixel Celsius temperatures, generates thermal orthophotos through an OpenSfM pipeline, and runs automated hotspot QC with auto-masking (YOLO-OBB, ONNX), feature extraction, and spatial filtering — all surfaced in a built-in map viewer.
End-to-end drone land survey across 70–80 km²: placed ground control points with static RTK measurements, planned and flew the survey missions, and processed the captured data into final orthophotos, DEMs, and mouza maps.
IP-camera-based attendance platform that detects and recognizes enrolled people using YOLO, InsightFace, and AdaFace, with custom algorithmic engineering for reliable presence tracking and a live management dashboard.
Multi-camera, multi-object tracking system for drone footage using YOLO and the VisDrone MOT dataset. Engineered data transformation scripts to convert VisDrone annotations into YOLO format for fine-tuning and synchronized tracking across camera feeds.
End-to-end data acquisition, training, and production pipeline around the open-source OmniCloudMask library for satellite cloud and shadow segmentation, fine-tuning baseline models on custom-curated datasets to boost accuracy for target use cases.
Implemented an instance segmentation architecture producing bounding boxes and masks, built on a Feature Pyramid Network (FPN) with a ResNet101 backbone.
Developed a geospatial machine learning pipeline utilizing CatBoost and XGBoost algorithms for accurate kelp detection in multi-spectral LANDSAT imagery.
Exploring content-aware strategies in deep learning-based change detection for robust bi-temporal satellite image analysis.
Full-stack job management app with JavaScript and MERN stack built from scratch. Robust authentication and efficient error handling. Add, edit, delete jobs with rich properties for easy filtering and search.
jawad@portfolio: ~/skills
~ $ ls -la skills/
~ $
[Oct 2023 - Present]
@ HawarIT Limited, Dhaka, Bangladesh
Architecting and shipping commercial AI systems in Python and C++ that turn raw aerial and satellite imagery into production products — spanning computer vision, large-scale data engineering, and photogrammetry. Key work includes AEROCheck (image quality control), building-footprint change detection, terabyte-scale Cloud Optimized GeoTIFF pipelines, containerized ML microservices, and seamline correction.
[Jun 2023 - Nov 2023]
@ Freelancing, Dhaka, Bangladesh
Delivered full-stack web applications to international clients — an NFT marketplace, a Bitcoin trading dashboard, and a buy-sell-rent platform — using React, Node.js, Express, and GraphQL. Also completed a medical data-annotation project preparing imaging data for ML training.
[Jan 2019 - May 2023]
@ Islamic University of Technology, Gazipur, Bangladesh
I graduated from Islamic University of Technology after 4 years of undergraduate studies in the department of Computer Science and Engineering, with a GPA of 3.7/4.0. My thesis explored ML-based data augmentation with semantic similarity for Bangla, a low-resource language.
[Nov 2021 - Dec 2021]
@ Battery Low Interactive Ltd., Dhaka, Bangladesh
I interned for Battery Low Interactive Ltd. in game development. As a final assessment my team and I submitted a complete game developed with Unity 2D game engine.
Got a project, a role, or just want to talk computer vision and geospatial AI? My inbox is always open — I'll get back to you as soon as I can.