GNN-Based Floor Plan Line Segment Parser (Bachelor's FYP) - Tech: Python · Graph Neural Networks · Computer Vision · Django · REST API · React (frontend)
• Built a machine learning model using Graph Neural Networks that could look at architectural floor plan images and automatically detect and classify the different lines in them, a computer vision problem.
• Built the full system end-to-end: a Django backend to run the model, plus a React frontend where users could upload a floor plan and see the results in real time.
• Got the whole pipeline working end-to-end, from a raw floor plan image all the way to the finished, classified output.
• Organized the code into clear, separate parts (model, API, and interface) so each piece could be built and tested on its own.
github.com/969Furqan/floorplan-front-end-and-backend
Preference-Oriented Fixed-Priority Scheduler - Tech: Python · Real-Time Scheduling · Algorithm Implementation · OS Concepts
• Built a task-scheduling algorithm from a research paper, which decides what order tasks should run in based on priority and preference rules.
• Turned the paper's mathematical proofs into working, tested code, showing I could read dense academic material and actually build from it.
• Checked my results against the paper's own test cases to confirm the scheduler was working correctly.
github.com/969Furqan/Preference-Oriented-Fixed-Priority-Scheduler
Freelance Web Scraping Pipelines - Tech: Python · Scrapy · BeautifulSoup · Pandas · Data Cleaning
• Built several data-collection tools for clients, automating the process of pulling information from different websites.
• Handled common scraping challenges like multi-page listings and anti-bot blocks, and delivered the data in ready-to-use formats (CSV, JSON).
• Delivered clean, organized datasets that clients used to train and test their machine learning models.
github.com/969Furqan/scraping-work
