Python

Overview Built and deployed a production-grade automated pipeline for processing banking transactions at City International Exchange (Kuwait). This system eliminated manual processing errors, ensured 100% regulatory compliance, and achieved 70% reduction in processing time while serving 20+ end users. Problem Statement Banking transaction processing relied heavily on manual data entry and spreadsheet-based workflows, introducing recurring errors and creating bottlenecks. Compliance tracking was fragmented, making regulatory audits time-consuming and error-prone. Solution & Approach Architected an end-to-end automated pipeline combining:...

Overview A complete, production-grade implementation of the Chord protocol for distributed hash tables. This project demonstrates deep understanding of distributed systems, peer-to-peer networks, and consistent hashing. Supports dynamic node join/leave operations with automatic ring stabilization. Problem Statement Traditional centralized systems don’t scale. Distributed hash tables (DHTs) enable decentralized data storage and lookup. Chord is a foundational algorithm that guarantees O(log N) lookup time in a P2P network of N nodes—critical for scalable, fault-tolerant systems....

Overview A comprehensive movie recommendation system leveraging collaborative filtering and content-based filtering techniques to provide personalized suggestions. The system analyzes user ratings and movie metadata to identify patterns and recommend films tailored to individual preferences. Problem Statement Users searching for new movies often face decision paralysis with thousands of options. A personalized recommendation system can reduce discovery friction by leveraging collective intelligence from user ratings and similarity between movies to surface relevant content....

Overview An end-to-end machine learning pipeline that predicts soccer match outcomes with 72% accuracy using advanced feature engineering, ensemble methods, and hyperparameter optimization. Benchmarked 7 classifiers across 10 dataset variants with rigorous cross-validation. Problem Statement Traditional sports prediction relies on domain expertise and intuition. A systematic ML approach can leverage historical match data to identify patterns and predict outcomes, enabling data-driven insights for analysts and sports enthusiasts. Solution & Approach Engineered a comprehensive pipeline combining:...