Who said what, and was it right
A speculative project on AI assurance for court interpreting, and two problems underneath it: putting each turn against the right speaker, and keeping an eag...
Posts by Tag
- ai 12
- agents 6
- method 6
- data 5
- mcp 3
- python 2
- infrastructure 2
- web 2
- traceability 2
- music 1
- security 1
- llm 1
- engineering 1
- speech 1
ai
Terraform from my phone
Two days directing Claude from my phone while it stood up an AWS organisation, a permissions-bounded identity for itself, and a working AI demo on top. Mostl...
An escalation ladder for LLM judgement
Batched mid-tier models make the calls and adversarial verifiers try to break them. I read the two places they disagreed – notes on a cheap architecture for ...
What National Anthems Quietly Stopped Singing
I scored all 195 national anthems on how warlike they are, twice – the full official text against the verses people actually sing. Most countries have quietl...
Analytical Writing as Applied AI: a Method, Built From a Phone in Japan
A repeatable rubric-scored method for analytical blog posts – and what it means that I built the whole thing mobile-only with Claude Code, on holiday in Japa...
IATI Traceability, Upgraded: Graph Database + MCP
Revisiting international aid traceability using a Neo4j graph database and Claude MCP — a fresh take on a problem I first tackled at the UK Department for In...
iTunes Wrapped: Elasticsearch + XML + Claude MCP
A day-build that parses an iTunes XML library, indexes it into Elasticsearch via Docker, and queries it using Claude MCP — a personal ‘Spotify Wrapped’ built...
A ‘Day-Build’ with Cursor and Typescript
Building a quantum-art visualisation in a single day using Cursor and TypeScript — an experiment in agentic front-end development and what it feels like to s...
Quick Reflection on SVGs AI Image Generation
Why using Claude to generate SVGs is more interesting than standard AI image generation — the output is editable, composable code, not a pixel blob.
Model Context Protocol and ‘Enveloping’ the world for AI
Exploring how MCP and ‘enveloping’ infrastructure empower AI agents more effectively than raw model capability alone — with practical examples from agentic P...
Historic Project: Using LLMs to get Metrics from Text
An early applied NLP project using LLMs to extract structured metrics from unstructured text — built with Python and Streamlit before the current wave of AI ...
agents
Terraform from my phone
Two days directing Claude from my phone while it stood up an AWS organisation, a permissions-bounded identity for itself, and a working AI demo on top. Mostl...
An escalation ladder for LLM judgement
Batched mid-tier models make the calls and adversarial verifiers try to break them. I read the two places they disagreed – notes on a cheap architecture for ...
Analytical Writing as Applied AI: a Method, Built From a Phone in Japan
A repeatable rubric-scored method for analytical blog posts – and what it means that I built the whole thing mobile-only with Claude Code, on holiday in Japa...
A ‘Day-Build’ with Cursor and Typescript
Building a quantum-art visualisation in a single day using Cursor and TypeScript — an experiment in agentic front-end development and what it feels like to s...
Model Context Protocol and ‘Enveloping’ the world for AI
Exploring how MCP and ‘enveloping’ infrastructure empower AI agents more effectively than raw model capability alone — with practical examples from agentic P...
method
Who said what, and was it right
A speculative project on AI assurance for court interpreting, and two problems underneath it: putting each turn against the right speaker, and keeping an eag...
Terraform from my phone
Two days directing Claude from my phone while it stood up an AWS organisation, a permissions-bounded identity for itself, and a working AI demo on top. Mostl...
An escalation ladder for LLM judgement
Batched mid-tier models make the calls and adversarial verifiers try to break them. I read the two places they disagreed – notes on a cheap architecture for ...
What National Anthems Quietly Stopped Singing
I scored all 195 national anthems on how warlike they are, twice – the full official text against the verses people actually sing. Most countries have quietl...
Analytical Writing as Applied AI: a Method, Built From a Phone in Japan
A repeatable rubric-scored method for analytical blog posts – and what it means that I built the whole thing mobile-only with Claude Code, on holiday in Japa...
data
What National Anthems Quietly Stopped Singing
I scored all 195 national anthems on how warlike they are, twice – the full official text against the verses people actually sing. Most countries have quietl...
Analytical Writing as Applied AI: a Method, Built From a Phone in Japan
A repeatable rubric-scored method for analytical blog posts – and what it means that I built the whole thing mobile-only with Claude Code, on holiday in Japa...
IATI Traceability, Upgraded: Graph Database + MCP
Revisiting international aid traceability using a Neo4j graph database and Claude MCP — a fresh take on a problem I first tackled at the UK Department for In...
iTunes Wrapped: Elasticsearch + XML + Claude MCP
A day-build that parses an iTunes XML library, indexes it into Elasticsearch via Docker, and queries it using Claude MCP — a personal ‘Spotify Wrapped’ built...
Historic Project: Using LLMs to get Metrics from Text
An early applied NLP project using LLMs to extract structured metrics from unstructured text — built with Python and Streamlit before the current wave of AI ...
mcp
IATI Traceability, Upgraded: Graph Database + MCP
Revisiting international aid traceability using a Neo4j graph database and Claude MCP — a fresh take on a problem I first tackled at the UK Department for In...
iTunes Wrapped: Elasticsearch + XML + Claude MCP
A day-build that parses an iTunes XML library, indexes it into Elasticsearch via Docker, and queries it using Claude MCP — a personal ‘Spotify Wrapped’ built...
Model Context Protocol and ‘Enveloping’ the world for AI
Exploring how MCP and ‘enveloping’ infrastructure empower AI agents more effectively than raw model capability alone — with practical examples from agentic P...
python
iTunes Wrapped: Elasticsearch + XML + Claude MCP
A day-build that parses an iTunes XML library, indexes it into Elasticsearch via Docker, and queries it using Claude MCP — a personal ‘Spotify Wrapped’ built...
Historic Project: Using LLMs to get Metrics from Text
An early applied NLP project using LLMs to extract structured metrics from unstructured text — built with Python and Streamlit before the current wave of AI ...
infrastructure
Terraform from my phone
Two days directing Claude from my phone while it stood up an AWS organisation, a permissions-bounded identity for itself, and a working AI demo on top. Mostl...
Model Context Protocol and ‘Enveloping’ the world for AI
Exploring how MCP and ‘enveloping’ infrastructure empower AI agents more effectively than raw model capability alone — with practical examples from agentic P...
web
Password-Protected Drafts on a Static Site: StatiCrypt, Jekyll, and Honest Security Testing
How I added AES-encrypted preview posts to a GitHub Pages blog — with a 13-vector security test to prove it actually holds up.
A ‘Day-Build’ with Cursor and Typescript
Building a quantum-art visualisation in a single day using Cursor and TypeScript — an experiment in agentic front-end development and what it feels like to s...
traceability
An escalation ladder for LLM judgement
Batched mid-tier models make the calls and adversarial verifiers try to break them. I read the two places they disagreed – notes on a cheap architecture for ...
IATI Traceability, Upgraded: Graph Database + MCP
Revisiting international aid traceability using a Neo4j graph database and Claude MCP — a fresh take on a problem I first tackled at the UK Department for In...
music
Using the Harmonic Sequencer on the Ableton Push
A short video walkthrough of the Harmonic Sequencer on the Ableton Push — made to share with Subsine Academy coursemates and hopefully useful to anyone who s...
security
Password-Protected Drafts on a Static Site: StatiCrypt, Jekyll, and Honest Security Testing
How I added AES-encrypted preview posts to a GitHub Pages blog — with a 13-vector security test to prove it actually holds up.
llm
Who said what, and was it right
A speculative project on AI assurance for court interpreting, and two problems underneath it: putting each turn against the right speaker, and keeping an eag...
engineering
Who said what, and was it right
A speculative project on AI assurance for court interpreting, and two problems underneath it: putting each turn against the right speaker, and keeping an eag...
speech
Who said what, and was it right
A speculative project on AI assurance for court interpreting, and two problems underneath it: putting each turn against the right speaker, and keeping an eag...