I'm Kasitphoom Thowongs - an MSc Advanced Computing student at Imperial College London. I graduated with First Class Honours from the University of Glasgow & KMITL. I work across the whole stack, from web products down to the infrastructure that runs them, and I like software that feels fast and holds up under real use.


I'm an MSc Advanced Computing student at Imperial College London. Before that, I studied Software Engineering jointly at the University of Glasgow and King Mongkut's Institute of Technology Ladkrabang (KMITL), and graduated with First Class Honours. My focus is full-stack web, particularly the bit where design and engineering meet.
I'm happiest writing Next.js apps with thoughtful animations, fast servers, and interfaces that feel a little alive. Lately I've been exploring edge runtimes, real-time collaboration, and design tooling, and I'm spending more time on DevOps: containerisation, CI/CD pipelines, and infrastructure as code. I'm also keen on agentic AI, and on using AI agents to optimise everyday workflows so people can spend less time on repetitive tasks and more on the work that matters.
Off-keyboard you'll find me playing drums with my band, reading liner notes, or chasing good light around London.

A low-code platform for personalised customer communications, built as my Honours Individual Project.
Individual Project - University Dissertation
UniComm is a low-code Customer Communications Management (CCM) platform that lets business users design PDF templates, manage customer data and run personalised campaigns without waiting on IT. It is a multi-tenant SaaS with database-per-tenant isolation, a drag-and-drop template builder and a worker-based engine for large batch PDF runs. Single PDFs render in about 1.6 seconds, and per-document cost stayed stable from 50 to 5,000 customers.
Offline-ready QGIS maps from natural language, built with Thales UK.
Lead developer · Electron, Next.js, GPT-4o-mini, QGIS
A proof-of-concept built with Thales UK that lets users describe an area in plain language and get an offline-ready QGIS map. GPT-4o-mini turns the request into Overpass QL queries, and the data is visualised in QGIS. I led the architecture and all client demonstrations for a 6-person team. The project won the Grow Prize (Best Idea) and the Level 3 Honours Team Project Prize.
AI that gives musicians songs, chords, and lyrics tailored to their taste.
Hackathon team of 5 · React, Express, OpenAI
A web app built at the GUTS DYHTG Hackathon 2024 for the guitarguitar challenge. It uses AI to recommend songs and generate chord progressions based on each musician's genre, skill level, mood and instrument. The project won 2nd place.
Full-stack developer in a 6-person team, building production-ready modules for e-commerce and CMS products, from the frontend through to the database.
Led a 6-person team building a proof-of-concept with Thales UK that turns natural language questions into offline-ready QGIS maps.
Built generative AI tools that track sales conversions from Messenger and Instagram chats, and fast e-commerce pages focused on Core Web Vitals.
Internships, freelance, or just a chat about Next.js, music, or great UI — my inbox is open.