Selected work
The projects that best show how older academic work turned into usable tools.
I’ve separated representative projects from the wider archive so the portfolio reads as a curated body of work rather than a raw coursework dump.
Selected
6
The strongest older projects to keep front and centre.
Archive
9
Older academic work kept online as supporting context.
Focus
62
Themes across modelling, infrastructure, and scientific computing.
Featured
These are the older projects worth showing immediately because they either have a live interface, a stronger methods story, or a clear through-line into your current work.
2026 • Independent
AnimStudio — AI Animation Pipeline
An early-stage pipeline for producing 2D anime-style YouTube content from plain-text screenplays — parsing character descriptions, generating storyboards, synthesising voice, applying lip sync, and assembling a final MP4.
2026 • Independent
cviz — 2.5D Codebase Visualizer
A native Rust + wgpu application that turns git repositories into interactive 2.5D spatial maps. Files cluster by co-change history and semantic similarity. A Unix socket hook lets Claude Code light up files in real time as the agent navigates.
2026 • Independent
DjangoShip — Django SaaS Boilerplate for Europe
A paid Django 5.2 SaaS boilerplate for European developers — auth, Stripe, team management, and GDPR compliance built in. Extracted from a live production codebase and sold on Gumroad.
2026 • Independent
Football Team Tracker — Budget GPS Analytics
A budget GPS and fitness tracking system for amateur football teams. Django dashboard with Leaflet maps and Chart.js trend charts. ESP32 T-Beam hardware loggers at ~€50/player vs €200–300 for commercial equivalents.
2026 • Independent
RoleUp — Irish Job Board with Recruiter Transparency
A live Irish job board with a public recruiter response scoring system, salary benchmarks from 22 data sources, and subscription data products. Built with Django 5.2 + HTMX.
2026 • Independent
Algorithmic Trading System — Personal Hedge Fund
A multi-strategy algorithmic trading system running in dry-run on a Mac Mini. Three parallel strategies, an XGBoost signal confidence filter that doubles win rate, and a delta-neutral carry trade layer. Full monitoring via TimescaleDB + Grafana + Telegram.
Archive
Older work, cleaned up and kept in context.
Archive entries stay accessible, but the framing is intentionally different: these are part of the trajectory, not the headline.
Academic Archive
2021 • Academic
Bitcoin Price Time Series Analysis
ARIMA and SARIMA modelling of Bitcoin price data, including stationarity testing with Augmented Dickey-Fuller, ACF/PACF analysis, and price forecasting.
2021 • Academic
Data Architecture & Real-Time Analytics
A data architecture assignment building a PostgreSQL database pipeline for real-time Bitcoin price analytics, combining SQL schema design, Python ETL, and time series queries.
2021 • Academic
Data Ethics Essays
Two ethics essays exploring the ethical implications of large-scale biometric data systems (Aadhaar) and a broader case study on data ethics in modern technology.
2021 • Academic
Football Match Result Prediction
Multi-class classification of Premier League match results (Home Win / Draw / Away Win) using Random Forest, KNN, SVM, AdaBoost, MLP, and Naive Bayes classifiers on 8 seasons of match data.
2021 • Academic
Genomics PCA Dashboard
The original interactive dashboard for the genomics PCA work, built as the first web interface for exploring population structure in the 1000 Genomes dataset.
2021 • Academic
Genomics PCA Analysis
A population-genetics project that started as academic PCA analysis and has since evolved into a public genomics explorer for browsing ancestry structure interactively.
2021 • Academic
Machine Learning Coursework Notebooks
A collection of Jupyter notebooks covering K-Means clustering, Random Forest classification on the Iris dataset, image classification with CNNs, and CA assignment work.
2021 • Academic
Statistical Analysis Coursework
Statistical analysis labs and assignments covering correlation analysis, hypothesis testing, regression, and exploratory data analysis using the mtcars dataset.
2018 • Academic
The Ising Model — Metropolis Monte Carlo
Monte Carlo simulation of the 2D and 3D Ising model using the Metropolis algorithm — computing energy, magnetisation, heat capacity, and susceptibility across a temperature sweep to find the ferromagnetic phase transition and Curie temperature.