Case studies
The decisions behind the code.
Projects show what I built. Case studies show how I think: what I tried, what I rejected, and why.
- 01ContextThe real problem
- 02DesignArchitecture and decisions
- 03ImplementationThe difficult parts
- 04ValidationMeasured results
- 05ReflectionWhat I learned
- Trustworthy AI retrieval from curated knowledgeHow can an AI system understand a natural-language reflection and retrieve real source material without inventing the material itself?MoodVersePublished
- Grounding AI recommendations in real page dataHow do you let an AI explain SEO issues and prioritise fixes without inventing problems the site doesn’t have?SEO MakerPublished
- Building an AI content intelligence pipelineHow do you turn content from many sources into relevant, searchable insights without blocking the application?Insight StreamPublished
- Building a reliable price-alert systemHow do scheduled data collection, threshold checks and notifications work together without duplicate or stale alerts?AlertChainPublished
- Designing secure project-level access controlHow do you make sure people only see the projects they are authorised to see, including from a separate AI service?WorkroomPublished