callumployed

Local-first job discovery and AI-powered application preparation.

• Built a local-first job-search platform with a web tracker, Python CLI, FastMCP server, and shared SQLite-backed data and service layers
• Automated job discovery with scheduled Playwright and Browserbase scans, dedicated ATS adapters, LangGraph orchestration, deterministic scoring, LLM fallback assessment, and schema.org JobPosting extraction
• Implemented company tiers, career links, role intake, configurable filters, notes, scan history, and application lifecycle workflows with durable status transitions
• Created an AI application workspace that indexes source materials and applicant details to generate, revise, and bulk-regenerate tailored resumes, cover letters, PDFs, application answers, and grounded role Q&A
• Developed a Firebase Central service for company reconciliation, passkey-gated private role feeds, privacy-safe scan telemetry, and an aggregate metrics dashboard while preserving local application data

See more here:  GitHub Repository. 

Find My Force

RF transmitter identification using learned and statistical signal features.

• Built an RF signal classification pipeline combining CNN-extracted features with statistical signal features for transmitter identification
• Achieved 89% classification accuracy by ensembling PyTorch CNN, XGBoost, and logistic regression models
• Engineered spectral features from RF data in the complex plane including phase jump frequency and amplitude variance
• Created for the Red Team Hacks Defence Hackathon

See more here:  GitHub Repository. 

TBI Motor Recovery

AI-assisted 3D motor-rehabilitation exercises for traumatic brain injury.

• Built an interactive 3D motor-rehabilitation platform using Three.js for individuals with traumatic brain injury to perform recovery drills
• Combined eye tracking and device-orientation data to measure patient movement
• Developed AI-generated rehabilitation exercises using patient motion data, PCA-based analysis, and real-time 3D visualization
• Visualized exercises in-browser with Three.js and Web Bluetooth for external controls
• Created for the 2026 Telus AI at the Edge of Innovation Hackathon

See more here:  GitHub Repository, Slide Deck. 

FITNET

Real-time wearable motion tracking with synchronized EMG and IMU data.

• Built a real-time wearable motion-tracking system using ESP32-S3 nodes, EMG, and 9-axis IMUs
• Achieved sub-millisecond synchronization across distributed sensors over Bluetooth Low Energy
• Achieved 95% movement-classification accuracy using DTW-KNN on synchronized EMG and motion data
• Implemented on-device Mahony Attitude and Heading Reference System (AHRS) for accurate orientation estimation
• Streamed and processed sensor data in a SwiftUI/SceneKit iOS app for real-time 3D motion visualization
• Created a custom PCB and power system in Autodesk Fusion 360 with battery protection and charging

See more here:  Firmware, iOS App, GitHub Organization. 

Drone from Scratch

A custom drone spanning hardware, embedded firmware, and mobile control.

• Utilized a variety of low-level communication protocols such as SPI, PWM, I2C
• Researched components based on electrical requirements (FPGA vs microprocessor, voltages, PWM channels, camera interface)
• Designed custom chassis and elecrical diagram with Autodesk Fusion 360
• Ran no-std Rust on microprocessors for memory safety and speed
• Designing flight controller and remote-drone commincation protocols
• Creating iOS app to connect via bluetooth to ground station which communicates with drone
• Read a variety of sensor feeds on an embedded device (accelerometer, barometer, magnetometer, camera, ultrasonic)

See more here:  CAD Model Link, GitHub Repository. 

VitAlert

A responsive health-information app built at nwHacks 2024.

• Submission for NWHacks 2024, a 24 hour hackathon focusing on health
• Worked on collaborative coding skills with 3 group members
• Developed React skills with a responsive UI that adapts to screen size
• Utilized R for webscraping and data tidying
• Linked in APIs from several services

See more here:  Live Demo, GitHub Repository. 

Virtual Closet

A fully tested Java desktop app for organizing outfits.

• Users organize closets, individual pieces of clothing, and outfits to track their style
• Practiced imperative UI development with Java Swing
• Leveraged Java's inherant OOP to enforce SOLID design principles
• Has 100% model LOC test coverage to ensure a bug-free user experience
• Created UML diagrams for both the model and user interface class structures
• Term project for CPSC 210 (Software Construction) at UBC

See more here:  Virtual Closet. 

Classification of Heart Disease

Heart-disease prediction using KNN and exploratory data analysis.

• Trained KNN classification model on whether patients are likely to have heart disease
• Used R and Jupyter notebook
• Created data visualizations for ease of understanding
• Performed exploratory data analysis (EDA) to give insights to variable selection
• Researched possible links between variables
• Created step-by-step report with rationale, and formed a conclusion based on data

See more here:  Report, GitHub Repository. 

Deloitte ThinkTECH

A secure medical-form workflow that placed top four at ThinkTECH.

• Created for the Deloitte ThinkTECH business case competition
• Improved technical communication by presenting to 100+ individuals at the Deloitte summit building in Vancouver
• Safely managed sensetive medical forms with cryptographic principles
• Integrated online form APIs to track information
• Set up webhooks to quickly notify and store input data
• Placed top 4 of 160+ candidates

Exvi Fitness

A serverless fitness tracker for Android and desktop.

• Serverless backend android and desktop fitness tracking app
• Used AWS for hosting, database, cloud functions, and a RESTful API
• Created user system from scratc with dynamic salting and hashing, and two factor authentication
• Implemented custom serialization scheme for quick and accurate data saving
• Leveraged Java and Kotlin together via JVM for cloud and client code

See more here:  Poster Report, Exvi. 

Rendering Engines

2D and 3D rendering engines built across four languages.

• A series of 3D & 2D rendering engines created from scratch
• Implemented the entire render pipeline to draw to a console window with ASCII
• Used OpenGL bindings in various languages to utilize GPU power
• Implemented in C++ (with CMake), Java, Rust (Desktop & WASM), and Typescript (WebGL)
• Used PBR techniques, a variety of light types, specular/diffuse/ambient lighting, and normal mapping
• Implemented linear algebra libraries for quick transformation calculations

See more here:  Java, C++ with physics & multiple platforms, C++ 2D & 3D with lua scripting, Typescript, Rust, Raytracer. 

Java

Typescript

C++ (text rendering)

Java with multiple render views

Typescript

C++ (OpenGL)

Nexus

A Windows utility for apps, wallpapers, and privacy settings.

• Created with Windows WPF, .NET, and C#
• A center for common actions which I used frequently such as wallpapers and apps
• Allowed animated wallpapers in the form of GIFs or MP4s to be set
• Let users launch single apps, groups, or all
• Allowed disabling of certain Microsoft tracking features

See more here:  GitHub Repository.