Most AI projects don't fail because of model choice — they fail because of dirty data and brittle pipelines. I build the boring infrastructure — proxy rotation, anti-detection scraping, ETL at scale — as carefully as the models sitting on top of it. That's the discipline I bring from data engineering into applied ML.
I'm a Software Engineering graduate (BS, CGPA 3.79/4.00) from The Islamia University of Bahawalpur. My final-year project combined Selenium/Scrapy scraping, ETL pipelines and Docker-deployed TensorFlow/PyTorch models behind a REST API — an AI-powered e-commerce recommendation system, end to end. That's still the template I work from: real data in, deployed system out.
Since 2022 I've freelanced on Fiverr while also serving as a Data Engineer at Code Apex, Islamabad. I'm currently exploring drift-aware cold-start recommendation systems as a research direction and looking to bring that same discipline to a full-time AI/ML or data engineering team or take it further at the graduate level.