AI/ML Engineer · Dublin, Ireland
Multi-agent systems, RAG infrastructure, and LLM applications that run in production.
LinkedIn · Email · All repositories
I build software and agentic systems end to end, designing them, building the orchestration layers, the retrieval layers underneath, the APIs that serve it, along with the people facing interfaces.
MSc Computer Science (AI/ML focus) from University College Dublin, First Class Honours, 2026. EU work authorised.
| Project | What it is | Stack |
|---|---|---|
| finance-autopilot | Six-agent adversarial pipeline that red-teams an investment thesis. A private-equity-style Critic and an Upside Modeler are held to opposing Pydantic constraints, so they are structurally incapable of agreeing. Includes a React Flow view of the agent graph and a live "Interrogation Room" chat. Built at a hackathon; it led directly to my internship offer. | LangGraph, Pydantic, FastAPI, Next.js, React Flow |
| glowai | Mobile-first skin and hair analysis web app. Vendor analysis sits behind a single adapter interface, so the mock and the live integration swap with one feature flag. | Next.js, TypeScript |
| baseline-guardv3 | GitHub Action that fails CI when JS/CSS uses web-platform features outside your Baseline target. Parses with Acorn and PostCSS, checks against the web-features dataset, and emits an HTML report. | JavaScript, GitHub Actions |
| MCP connector for a finance platform (private) | Model Context Protocol server that lets Claude query live platform data through a read-only FastAPI tool surface. | Python, FastAPI, MCP |
Representational Laundering: an essay on why connectionist models keep reinventing physiognomy. It uses a 2020 Nature Communications paper, which claimed to read trustworthiness from faces in historical paintings, to show how a modeller's assumptions come back out of a network looking like findings.
| Role | Company | Focus |
|---|---|---|
| AI/ML Engineering Intern | Assiduous (NovaUCD, Dublin) | AI-powered corporate finance platform for SMEs. Worked across the Next.js frontend and Python/FastAPI backend: agent tooling, MCP connector, React Flow graph editors, audit-trail event logging. |
| Lead AI/ML Engineer | Parbat (Pegasus Angel Accelerator) | Led AI/ML engineering at an Irish retail AI startup. |
| Data Engineering | Daraz (Alibaba Group) | Data engineering for a large e-commerce marketplace. |
| AI Engineering | Qavi Tech | Applied AI engineering. |
| Software Engineering Intern | KCompute | Software engineering. |
Agents and LLMs: LangGraph · LangChain · Anthropic API · Model Context Protocol · Pydantic
Retrieval: pgvector · PostgreSQL · RAG pipeline design
Backend: Python · FastAPI · Docker
Frontend: Next.js · React · TypeScript · React Flow
| Top 2% | AI Mathematical Olympiad (AIMO), Kaggle |
| 1st place | TiE Unicon, among 1,200+ teams |
| First Class Honours | MSc Computer Science, University College Dublin |
| Founder | FAST Entrepreneurship Society, FAST NUCES |
Open to AI/ML and Software engineering roles across Europe.