Projects and case studies
Five engineering projects spanning AI/ML, research, software, and Android delivery. Each case study explains the problem, my role, architecture, implementation, results, and lessons.
AI/ML Engineer · Software Engineer · TA / Instructor
I developed this PII-aware named-entity recognition evaluation framework for my completed M.A. thesis at York University.
My role: I designed and built the research pipeline, corpus standardization, and evaluation framework.
AI/ML Engineer · Software Engineer
An independently built, end-to-end hands-on RAG engineering project for grounded retrieval, citation-backed answers, strict refusals, evaluation, testing, and Docker delivery.
My role: I independently built the complete ingestion, retrieval, generation, evaluation, testing, and deployment workflow.
AI/ML Engineer · Android Developer · TA / Instructor
An assistive smart-glasses system combining Android, embedded hardware, computer vision, OCR, APIs, and text-to-speech.
My role: I delivered more than 80% of the system and built its Egyptian-currency recognition capability from scratch.
AI/ML Engineer · Software Engineer · Android Developer
A client-owned Android and machine-learning project for recreational scuba planning, safety-profile classification, and recommendations.
My role: I owned and implemented the project end to end, from the Android client and service integrations to the safety-classification workflow.
Software Engineer · Android Developer
An on-demand courier Android application with order workflows, remote and local data layers, notifications, and map-based tracking.
My role: I delivered the Android application and its mobile architecture and service integrations.
Additional work
Client applications, enterprise systems, research, and applied machine-learning delivery.
AI/ML · Research
A privacy-safe continual-learning method designed to reduce catastrophic forgetting without retaining raw historical PII.
AI/ML · Client delivery
More than ten applied projects across forecasting, segmentation, recommendation, sentiment analysis, fraud detection, and image classification.
Software · Enterprise
Document intake, review, routing, archival, retention, metadata validation, and operational support in OpenText Documentum.
Software · Backend
Java utilities for automated PDF generation, assembly, transformation, packaging, logging, and failure diagnostics.
Software · APIs
Java and Python services for prototypes and proofs of concept with clear JSON contracts, authentication, and error handling.
Android · Client project
A football-talent platform where players present their skills so club agents can scout and sign them.
Android · Client project
A private mental-wellbeing companion application with secure check-ins and local note storage.
Android · Client project
Recipe and meal-planning workflows with schedules, shopping lists, networking, and offline storage.
Android · Game
A casual action game developed with attention to rendering performance, controls, and input latency.
Android · E-commerce
A mobile commerce application with catalogue browsing, cart and checkout flows, and multi-currency support.
Android · Utility
Geolocated hourly and daily forecasts with caching and graceful network-error handling.
Android · Discovery
Restaurant discovery with map and list views, filtering, saved state, and offline recent results.