Dafydd-Rhys Maund

Master's degree in Data Science, completed with Distinction. The course built on my software-engineering background with modules across big data, data mining, data visualisation, machine learning, computer vision, deep learning, optimisation, modelling, verification, operating systems, architectures and research methods.

My thesis focused on a hybrid machine-learning and rule-based support chatbot for university students, staff and visitors. I built the system end-to-end using Python, FastAPI, Rasa, PostgreSQL and a RoBERTa-based intent classifier.

The architecture combined ML classification for flexible intent detection with rule-based dialogue flows for parts of the system that needed tighter control, including escalation, ticket submission and sensitive-topic handling. The result was a modular support system with separable frontend, API, conversational logic and data-storage layers.

First Class Honours in Software Engineering, awarded the Excellence Scholarship. The degree gave me a broad software engineering foundation across programming, algorithms, databases, concurrency, declarative programming, web services, software testing, project planning, mobile apps, embedded systems, graphics and games programming.

Final-year dissertation: a real-time synchronised media-streaming application where users could watch, listen and chat together across YouTube, SoundCloud, Twitch and Facebook in shared rooms.

Built with React, Node.js and Socket.io, the project handled persistent rooms, playback synchronisation, drift correction and live chat. It was a proper full-stack build: frontend state, backend events, real-time communication, third-party media integrations and enough edge cases to make timing bugs feel personal.

The course also developed the less glamorous but important parts of engineering: requirements analysis, project specification, documentation, testing, planning, teamwork and turning a rough idea into something that can actually be delivered.

Open-source contributions to PHPOffice/PHPWord and other libraries I use or rely on. Mostly the kind of work that looks small from the outside but matters in real projects: bug fixes, documentation improvements, examples, edge-case investigation and clearer reproduction steps.

The next step is to contribute more deliberately across the stack I work with: PHP tooling, backend frameworks, API libraries, documentation, testing utilities and data/ML projects connected to my master's work.

My focus is useful maintenance over vanity contributions: failing test cases, clearer docs, small bug fixes, migration notes, typed interfaces, validation edge cases and anything that makes a project easier for the next developer to use.