All Projects

Volunteered to build a self-improving sentiment classification web app for Matters, a Hong Kong publisher. The model updated with each user interaction.

First LLM classification model using Hugging Face BERT, developed at Microsoft. Initial hands-on foray into state-of-the-art NLP.

Composed a suite of introductory NLP tutorials covering tokenisation, stemming, lemmatisation, transformers, and attention mechanisms. Published in Towards AI and Towards Data Science.

Traditional NLP sentiment model trained on L.A. restaurant reviews.

NLP sentiment analysis model applied to New York restaurant reviews.

Social listening project using NLP to analyse sentiment around McDonald's in London.

Topic modelling and collaborative filtering system for action game recommendations.

Kaggle competition. LLM-based text quality evaluation. Top 32% out of 4,700+ teams.

Kaggle competition. Multiple-choice science exam solved by LLM. Top 42% out of 4,700+ teams.

In-depth Kaggle tutorial on 4-bit quantisation and its impact on Llama2 performance.

Designed and trained a small language model end-to-end, sharing the process and insights with the community.

Chatbot with the personality and vocabulary of a 5-year-old, built with LangChain and OpenAI.

LLM-powered tool for generating sales copy, tweets, and product descriptions for different target audiences.

Conversational chatbot capable of recalling previous parts of the conversation, such as a user's name.

Article dissecting the mathematics of positional embeddings in NLP models and their integration with original embeddings, using drawings to make it accessible.

Self-initiated pilot exploring a local LLM for council email handling, presented as an AI project proposal to Wandsworth Borough.

Real-world use-case article and code: building an AI-powered research assistant with a modern RAG stack.

Earned the Fundamentals of Azure AI Services badge in the AI Skills Challenge.

Completed the "Operationalizing LLMs on Azure" certificate course by Duke University.

Awarded the AI Skills Challenge badge from Microsoft Azure.

Kaggle competition. Mathematical reasoning with LLMs at olympiad level. Highest personal ranking.

End-to-end series covering motivations, machine setup on WSL2, and results from fine-tuning open-source LLMs for Text-to-SQL tasks.

AI-powered registration system for local government. FastAPI backend deployed on Render, Gradio frontend on HuggingFace.

Fine-tuned GPT-OSS 20B into a distinct debate personality via QLoRA on RTX 4090. Full pipeline: dataset curation, fine-tuning, Transformers deployment.

Hands-on guide fine-tuning OpenAI's GPT-OSS 20B model for multilingual reasoning on an RTX 4090.

Tutorial on building a composable AI agent using the Model Context Protocol (MCP) for stock investment analysis.

Fine-tuned GPT-OSS 20B on PubMedQA with chain-of-thought reasoning, achieving 73.6% accuracy — rank 19th on the PubMedQA leaderboard. Trained via QLoRA on RTX 4090.

Optimised Google's Gemma-3-270M via structured pruning and knowledge distillation — exploring model compression techniques for efficient LLM deployment.

AI chatbots with distinct characters and faces — a product exploring personality-driven LLM interactions.

Built a personal automation app using OpenClaw on macOS — AI-powered inbox management.

Multi-agent RAG chatbot over 1.76M HM Land Registry transactions. 16 years of data, all 33 London boroughs, chart output on every answer. Flagship project.

Automated monthly data pipeline publishing updated Kaggle datasets and mini-reports from HM Land Registry Price Paid data. Runs every 5th of the month.