Waheed Abbas

Led a cross-vertical recommendation engine across 5-variant A/B tests, driving $1M incremental order value and a 13.2% YoY conversion lift. Shipped multiple ML initiatives including location-of-interest modeling, an LLM email/WhatsApp parser at 99% accuracy/98% automation, and learning-to-rank hotel re-ranking (+69% conversion, +43% GBV).

Built an implicit-feedback collaborative filtering recommender, improving conversion by 7% and speeding time-to-discovery by ~1 minute per session. Developed an order-delay prediction system (81% precision/79% recall) to reduce cancellation rate by 3% and lower monthly lost-order costs.

Built and deployed a deep-learning video summarization tool that condenses long videos into 30-second highlight clips. Developed a pose-estimation-based basketball performance analysis system (>90% accuracy) and built prediction/optimization models for highway accidents (70% accuracy) and scooter dock placement (~7% profit gain in affected zones).

Built and delivered a content-based job and course recommender end-to-end, including data collection, a Flask API, and the recommendation engine. Developed an abusive-content detector (94% accuracy) and a time-series demand-forecasting system that reduced inventory carrying costs for a retail client.

Co-built a question-bank platform for the Saudi Health Commission using PHP/CodeIgniter, supporting thousands of daily users, and developed a backend/front-end currency-exchange dashboard. Applied data-driven modeling to optimize scooter dock placement (~7% profit gain in affected zones) and built a highway-accident prediction system.

Grade: CGPA: 3.48 / 4.00

Activities and societies: Merit-based scholarship each semester.

Earned an M.S. in Computer Science from COMSATS University Islamabad. Achieved a CGPA of 3.48/4.00 and received merit-based scholarship each semester.

Completed a B.S. in Computer Science at Lancaster University. Studied foundational and advanced computing concepts in preparation for graduate-level work.