I build AI systems that solve real problems — from government MIS platforms to published ML research. Self-taught. Based in Pakistan. Currently a Research Assistant at DIP Lab & BS student at Agriculture University of Peshawar.
I'm Muhammad Haris Afridi — a passionate, self-taught AI Engineer and Full-Stack Developer from Peshawar, Pakistan. I build practical tools and data-driven applications, including government systems and ML projects, with a focus on solving real problems for underserved communities.
Pursuing a Bachelor of Science in Artificial Intelligence (2023–2027) at Agriculture University of Peshawar. Simultaneously serving as a Research Assistant at the Digital Image Processing (DIP) Lab, Islamia College Peshawar — where I work on applied computer vision research.
Modern restaurant platform featuring online ordering, menu management, automatic order tracking, dine-in POS, cash payments, admin dashboard, and thermal receipt printing.
Modern computer-store platform featuring product catalogs, categories, search, stock management, product details, shopping cart, WhatsApp ordering, and an admin dashboard.
Independently developed full-stack platform for local government planning and management, featuring intelligent project prioritization, multi-role access, reporting, and workflow management.
N8N workflow pipeline for automated AI content generation — content brief triggers GPT-powered script generation, formats structured output, and delivers ready-to-use content. Built for social impact topics including health awareness and community campaigns.
Benchmarked 13 ML models on 58,750 PakWheels listings for used car price prediction in Pakistan. LightGBM achieved R² = 0.9676. Introduced domain-specific features (vehicle generation encoding, brand origin, city tier, trim grade) validated via SHAP analysis. Deployed as a public Streamlit app. Manuscript under review at International Journal of Data Science & Analytics.