Ongoing · product development
ODIN — Offline Data & Intelligence Network
I’m developing ODIN, a privacy-first local AI desktop runtime bringing local models, hardware-aware routing and persistent memory into a controlled workspace.
Tauri · React / TypeScript · Rust · SQLite · Ollama
Project details
My design includes resource budgets, verification, telemetry and quantized model support. I’m working on product development and commercial planning.
Ongoing · foundation stage
VEID — View Everything In Detail
I’m building VEID, a staged educational-content production system with a modular Python foundation.
Python · Configuration · Rendering pipeline
Project details
I have completed the repository foundation. Planned components include Manim rendering, voice generation, captions and FFmpeg assembly; those integrations are roadmap items, not completed capabilities.
Ongoing · foundation stage
AETERRA — Planet Evolution Simulation
I’m developing AETERRA in Godot 4 to explore a living world through planet, environment, climate and eventual ecosystem systems.
Godot 4 · Procedural generation · Simulation architecture
Project details
I’m starting with planet foundation and camera controls. Evolution, ecosystems and multiplayer are longer-term goals rather than delivered features.
Completed project
Rust tower defence dissertation
I developed a real-time two-player tower defence game in Rust for my MSc dissertation.
Rust · Real-time systems · Game development
Explore project details
Context
I developed a real-time two-player tower defence game in Rust for my MSc dissertation.
Engineering focus
The project brought together gameplay state, unit spawning, movement, player actions and animation within a real-time game loop.
What I learned
I developed my understanding of systems programming, timing-sensitive behaviour, debugging and performance-aware design. Rust made ownership and memory management an important part of the implementation.
Completed project
Demand forecasting
I worked on Random Forest demand forecasting using historical business data.
Python · Random Forest · Forecasting
Explore project details
Context
I worked on demand forecasting at YCH Logistics for Lenovo PCG and DCG, connecting historical operational data with planning needs.
Approach
I used Python and Random Forest modelling to explore demand patterns in historical business data. My wider YCH work also involved operational data preparation, validation, reporting and Excel VBA automation.
Business focus
I focused on making operational information useful for planning and reporting. This project brought together business context, data quality and applied prediction.
Further detail
Model evaluation results, dataset definitions and implementation artefacts are not included in this public case study.
Live project
TechPathReady
I built TechPathReady to bring role-based roadmaps, skill assessments and browser-saved readiness reports into one experience.
HTML · CSS · JavaScript · localStorage
Explore project details
What I built
I built a static HTML, CSS and JavaScript learning platform with career roadmaps, skill quizzes and readiness reporting.
Assessment workflow
Learners choose a country and role, explore weighted skills and required topics, then complete topic-based tests. Answers are shuffled, and full test completion is required before scores are saved.
Results and learning
I added browser-local saved scores, a readiness dashboard, standalone practice quizzes and learning pages for Python, JavaScript, SQL, HTML, CSS and Java. The platform includes five scored career paths and a separate GCSE Maths and Python support path.
Engineering choices
I used localStorage for browser-local persistence and separated question banks, career data, scoring and presentation. Beta enquiries and feedback use external forms.
Explore the websiteCompleted project
Energy-efficient deep learning
I worked on deep learning model compression, exploring methods for making neural networks more efficient.
Quantization · Pruning · Knowledge distillation
Explore project details
Focus
I explored deep learning model compression and the trade-offs between model efficiency and predictive performance.
Methods
My work covered pruning, quantization and knowledge distillation: reducing model structure, lowering numerical precision and transferring knowledge into smaller models.
Technical perspective
The project strengthened my understanding of model evaluation and the relationship between accuracy, model size and inference efficiency. Specific benchmark values are not included here.
Completed project
Database design & SQL
I worked on relational modelling and SQL projects, using ER modelling, normalisation and database design.
SQL · ER modelling · Normalisation
Explore project details
Focus
I worked on relational database design and SQL, turning structured requirements into schemas and queries.
Methods
I used entity–relationship modelling, normalisation and database design to reason about relationships and data structure.
Skills developed
The work strengthened my understanding of relational modelling, query construction and preparing structured data for analysis.