Muhammad Abubakar

Masterarbeit Thema: Sorting of the single molecular FRET trajectories using Machine learning

Built and automated multi-client ETL pipelines ingesting data from OwnCloud via CyberDuck WebDAV API, performing data cleaning, preprocessing, and normalization before loading into PostgreSQL, following a Data Mesh architecture for domain-oriented data ownership. Deployed pipelines as cron jobs on OpenStack-provisioned VMs configured via Ansible Playbooks as Infrastructure as Code (IaC), and developed client-facing R Shiny app and Metabase dashboards for self-serve data querying and visualization.

Developed a Flask-based REST API backend and end-to-end ETL pipeline ingesting weather data from DWD and Meteostat, staging raw files into a Data Lake and loading into AWS Aurora Data Warehouse. Engineered preprocessing and feature extraction workflows correlating weather patterns with clothing inventory pricing insights, orchestrated via Celery task queues delivering analytics-ready datasets for SQL querying and Power BI visualization.

Working on computer vision, OCR (Tesseract) based project. Designing an algorithm to extract the text data from Images. Responsible to create the grids on the tables of invoices to extract data in an organized manner. Implementation of a transfer learning algorithm ( YOLO ) on the invoices to extract data. Implementation of multiple methods of OpenCV to enhance and tackle the problematic images. Generation of dictionaries and then conversion of dictionaries to the JSON format

Studied the approaches to read the human X-rays Worked on the development of real-time Machine learning system to detect the Pneumonia and Tuberculosis X-ray Applied multiple algorithms and compared their accuracies Used transfer learning and custom CNN Collected data from multiple platforms and the total count is 5556 Images Accuracy of the best model is 92.97% Deployed the model on Flask python framework Also wrote a research article which is currently on Arxiv.com

Worked with PHP framework Laravel Worked with CSS, Bootstrap , Html5, and Vanilla JS Responsible for creating UI and UX Integrated the 3rd party API's in the project for better UI/UX Responsible to create, design and normalize the database in SQL Also created the Docker container package of full-fledge IOT Django based application heavily based on the MS SQL Server, Redis