Practical technology, built where medicine meets software.

I'm Steve — a GP, software developer, and AI enthusiast. My work sits at the intersection of healthcare, artificial intelligence, and open-source technology. As a clinician I understand the pressures of the consulting room; as a developer I enjoy building the tools that ease them.

This is a small, scannable index of what I've built. Six sections, one page per project. Read what's relevant.

Over the years I've built projects ranging from medical education platforms and AI-powered clinical tools to self-hosted cloud infrastructure, conversational AI systems, automation platforms, and hardware home-lab environments. I like taking an idea all the way — combining software, networking, cloud, and modern AI into complete, working systems.

What drives me is curiosity: learning how systems work, solving hard technical problems, and sharing what I find through projects others can use, learn from, or build upon. Some began as answers to everyday clinical problems; others as excuses to explore an emerging technology — but all reflect a belief that thoughtful engineering can improve the way people work, learn, and live.

• 01 Search Agent → An agentic workspace in the browser. CrewAI multi-agent teams and a code-writing Smolagent, with live-streaming progress, provider fallback, a ~1,900-skill library, and one-click .docx export.
  • Python
  • FastAPI
  • CrewAI
  • Smolagents
  • LLM agents
• 02 CAD Studio → Describe a part in chat; Claude designs it with CadQuery, and you preview and download the STL from any device on your LAN. Live status streaming, an interactive three.js viewer, and conversational iteration.
  • Python
  • CadQuery
  • three.js
  • SSE
  • 3D printing
• 03 Scribe → An AI medical scribe. Capture a consultation from the mic, transcribe it with Whisper, and turn it into a structured clinical note — then fact-check every claim against the transcript in a second verification pass.
  • Node.js
  • Express
  • Whisper
  • Clinical NLP

• 01AI & Large Language Models
• 02Healthcare tech & clinical decision support
• 03Full-stack software development
• 04Self-hosted infrastructure, Docker & virtualization
• 05Home-lab engineering & automation
• 06Raspberry Pi, embedded & edge computing
• 073D design & additive manufacturing
• 08Open-source technologies

• 01 What actually pays → I pointed my own multi-agent Search Agent at a personal question — which flexible side hustles genuinely pay for a GP in the North East of England? — and kept only what survived fact-checking. A field guide to telemedicine, AI training, private clinics and the myths in between.
  • GP economics
  • North East
  • AI research
  • Side hustles
• 02 Your first AI app, built on a phone → You don't need an expensive laptop to build AI software. A walkthrough for doctors with zero coding background — build a working voice-transcription app on an Android phone using Termux, Node.js, an AI coding assistant and OpenAI Whisper, then finally settle the "why do I need an API key if I pay for ChatGPT?" question.
  • Getting started
  • Termux
  • Node.js
  • Whisper
• 03 Ten beginner AI projects, in order → One project is a spark; ten in the right sequence is a skill. A learning path of medical AI apps built back-to-back — each adding one new concept while reinforcing the last, ending in a full clinical documentation assistant. With the exact prompts and a Generate → Review → Explain workflow.
  • Learning path
  • Node.js
  • OpenAI
  • Prompts