Projects

Work worth a closer look

Featured projects first — production product work, autonomous systems, and hardware-adjacent software — then additional academic and ML projects.

Featured

  • Infection Prevention Compliance SaaS

    Full-stack development · IPAC Consulting

    Led product and full-stack development of a multi-tenant healthcare SaaS for infection-prevention and regulatory compliance workflows (Vue 3, Express/TypeScript, Prisma, MariaDB). Launched major features end-to-end including new database models, REST APIs, and migrations, while strengthening multi-tenant access controls and RBAC; contributed to an OpenAI RAG assistant over tenant knowledge libraries; owned weekly production deployments and code review for a 3-engineer team.

    • Vue 3
    • TypeScript
    • Express
    • Prisma
    • MariaDB
    • OpenAI

  • Queen's Autodrive — Behaviour Planner

    Behaviour Planner Software Lead · Autonomy Team

    Designed a behaviour planner in MATLAB Stateflow to drive an autonomous vehicle from map and perception inputs. Used MATLAB Coder to auto-generate C++ and verified the generated code against the Stateflow reference via SIL/PIL, plus RoadRunner simulation scenarios. Integrated planner behaviour with shared Autonomy Team interfaces in Git and presented verification progress at Autodrive executive design reviews.

    • MATLAB Stateflow
    • MATLAB Coder
    • C++
    • RoadRunner
    • SIL/PIL

  • Ku-Band Radio Telescope

    Software Lead · Queen's University Engineering Physics Capstone

    Led Python software for dual-axis stepper control over a serial interface, celestial pointing, and SDR sensor acquisition for a Ku-band radio telescope. Added command validation to reject unsafe zero/fractional motor steps; integrated motor control with intensity/PSD pipelines to build sky heat maps. Built OOP control software to track motor pose, pointing angles, and radiation readings for multi-hour unattended operation.

    • Python
    • PySerial
    • SDR
    • OOP
    View on GitHub

Also notable

Course and research projects that round out ML, simulation, and scientific computing experience.

  • Phishing URL Detection CNN

    Model design & training · Queen's University

    Built and trained a character-level CNN in PyTorch to classify phishing vs legitimate URLs from the URL string alone. Owned model design, training, and evaluation on ~190,000 examples, improving test accuracy from ~82% to 94.69% (96.06% phishing recall).

    • Python
    • PyTorch
    View on GitHub

  • Photonic Crystal Ring Resonator Thesis

    Engineering Physics Thesis · Queen's University

    Simulated coupled-cavity photonic crystal ring resonators to optimize frequency spacing, quality factor, and nonlinear mode overlap for SFWM-based squeezed light. Best configuration: 24-cavity ring with 6.31% nonlinear mode overlap.

    • Lumerical
    • MATLAB
    • Python
    View on GitHub

  • Computer Vision with Deep Learning

    Queen's University coursework

    Implemented and trained PyTorch models for image classification, object detection, and localization. Fine-tuned ResNet18 for vehicle presence detection at 96.2% test accuracy and built custom dataset loaders and evaluation tooling.

    • Python
    • PyTorch
    • ResNet18