Ten years of building through community projects, research and industry, with a Bachelors and Masters in Computer Science from Trinity College Dublin. I focus on dependable software that is useful, maintainable and straightforward to run.
Custom built affordable Software offerings for growing and mid sized companies
A private workspace that finds public buying and expansion signals, verifies each company, locates a real contact route, and scores the opportunity with AI. It can draft a first contact message and assist with the company’s own form, but it never sends or submits anything for you.
Open Source Free Product Offerings to give back to the community <3
Kinoir turns a folder of local movies and shows into a polished, private streaming library. It cleans messy filenames, finds metadata, and creates hover previews on your Mac. Grounded Ollama discovery and Kinoir Air keep exploration and home network streaming private, open source, native to macOS, and free from accounts.
Visit Kinoir
Pit Crew is a live BALL x PIT companion for faster decisions during a run. Track characters, balls, passives, levels, and available evolutions, then compare the upgrades the game offers with your current build. Recommendations respond to what you already have so the next useful choice stays easy to see.
Visit Pit CrewHover a logo to learn more.
Selected work across research, ML systems, and applied engineering.
Clinical ML early-warning system on MIMIC-IV ICU data, evaluated for deployment behaviour.
Multi-agent negotiation simulation with feedback-driven strategy optimisation, accepted at COLM 2025.
A CLI that deploys and manages AWS infrastructure from a handful of configuration lines.
Document classification plus regression to predict whether US state bills pass.
A side-scrolling town where each building is a project, with in-world product demos and sprites.
Across industry and research.
Co-authored a COLM 2025 accepted multi-agent simulation paper and won the Q4 Innovation title for a financial News-API demo, while building and fine-tuning sparse ML models and LLM workflows for financial signal extraction.
Built and optimised internal Oracle APEX applications and Essbase reporting workflows, including a JavaScript org-chart tool adopted across the team.
Built Cloud Builder, a CLI that deploys AWS infrastructure from compact config. It was selected for the Software Engineering Industry Awards and later open-sourced.
Trained an ML model to generate Figma UI code from design datasets, integrated the ChatGPT API into a client-facing product, and built an internal ML service with a custom API layer.
Built a text classifier over 1,000+ US legal documents and combined classification with regression to predict bill outcomes, migrating modeling to XGBoost and raising accuracy from 45% to 65%+.
Chaired Trinity's Network Society after serving as Events Officer, running technical community events across the 2021-2023 academic years.
Coordinated competitive programming events for Trinity's computer science community.
Research internship under Prof. Shobha Bagai, building a Java chess-simulation GUI and a modified variant with a decision-based opponent for probability experiments.
Led a school technology club after serving as Robotics Head, building battle robots and line-followers and running the annual tech symposium and online cryptic hunt.
Where the foundation was built.
Integrated B.A. (Mod.) & MSc Computer Science
DistinctionA combined undergraduate and master's degree in computer science, completed with overall distinction.
Secondary & Senior Secondary Education
93% in Class X national boards, 95% in Class XII national boards.
Peer-reviewed, thesis, and preprint work.
Self-optimising agents in a multi-agent social simulation.
COLM 2025
Threshold calibration and deployment behaviour on MIMIC-IV ICU data.
Master's thesis · Trinity College Dublin
Emoji sentiment stability research across 16,575 tweets.
Preprint · submitted to ICWSM 2027
The complete set, from flagship work to experiments and community projects.
Clinical ML early-warning system on MIMIC-IV ICU data, evaluated for deployment behaviour.
AI/ML GitHub → FlagshipMulti-Agent AIMulti-agent negotiation simulation with feedback-driven strategy optimisation, accepted at COLM 2025.
NLP GitHub → FlagshipCloud ToolingA CLI that deploys and manages AWS infrastructure from a handful of configuration lines.
Systems GitHub → FlagshipLegal-Tech MLDocument classification plus regression to predict whether US state bills pass.
AI/ML GitHub → Local-First AI Desktop AppLocal-first desktop AI app with multi-persona chat, on-device voice cloning, web retrieval, and GPU coordination.
AI/ML GitHub → FlagshipAnimated 2D WorldA side-scrolling town where each building is a project, with in-world product demos and sprites.
Web Open project → Agentic Development EnvironmentA personal fork of Emdash for coordinating multiple coding agents through isolated Git worktrees.
Systems GitHub →Emoji sentiment stability across 16,575 tweets, separating topical coherence from affective consistency.
NLPAn interpretability pipeline around LoRA adapters and sparse autoencoders.
AI/ML GitHub →Privacy-first Vedic astrology that computes sidereal charts entirely in the browser.
WebA marketing site for E.T.S Consulting Services with a custom design system and serverless contact flow.
WebAn ML prototype for design-to-code generation from proprietary design datasets.
AI/MLA configurable Java GUI for modified-rule chess and decision-tree experiments.
Early Work GitHub → Early ML ExperimentAn early Python face-recognition workflow with a MySQL-backed setup.
AI/ML GitHub → Language Tooling · Early WorkA Google Cloud experiment that transcribes Hindi video and extracts its subject matter.
NLP GitHub →Chaired Trinity College Dublin's Network Society after serving as Events Officer, running technical community events across 2021-2023.
2021-2023 LeadershipLed a school technology club after serving as Robotics Head, covering battle robots, line-followers, and the annual tech symposium and cryptic hunt.
2016-2020 LeadershipNo projects in this category.
Grouped by area, with the projects where each was applied.
Modeling, calibration, and evaluation built for deployment.
Retrieval, fine-tuning, embeddings, and multi-agent systems.
Shipping the systems around the model, end to end.
Experiment design and writing that holds up to review.
If you have a practical software problem, a product idea or an open source collaboration in mind, I am happy to talk. I work across software engineering, data and applied AI, from the first decisions through to dependable delivery.
