Multi-agent workflow & orchestration design
Subagents, fan-out/fan-in execution, plugin architecture and branch-per-phase Git strategies — designed so parallel agents stay traceable and reviewable.
Claude Code Developer & Architect
Developer and architect specialising in Claude Code and AI-augmented engineering.
Multi-agent workflows, structured analysis and design pipelines, full-stack apps. An enterprise background in trading, treasury and healthcare brings architectural rigour to AI-assisted development.
18 yrs
Enterprise delivery
10+
Agents orchestrated
2
Live frameworks & builds
01 — Services
Three ways I work with teams building on Claude Code.
Skills, hooks & MCP integration
Custom SKILL.md authoring and MCP servers that connect Claude Code to the external APIs and data your business already runs on.
Structured analysis, design & full-stack delivery
AORDL requirements engineering with full traceability, then delivery in Flutter Web, Go, Node and TypeScript — including RAG applications.
02 — Work
ROME
Requirements-Oriented Multi-agent Engineering
A methodology framework that orchestrates specialised Claude Code agents across a full SDLC pipeline — requirements through to build, with quality gates between every stage.
Read the framework page →- 10+ coordinated agents across the pipeline
- AORDL requirements language — 13 mandatory fields per requirement
- 40+ reusable skills
- Fan-out/fan-in parallel execution with PR-based quality gates
- Full forward and backward traceability
- Branch-per-phase GitHub strategy
Independent · 2024 — present
Florence
NHS management system
A scalable healthcare architecture with a Flutter Web frontend and Golang / Node.js backends. Claude Code drives rapid prototyping; RAG applications handle data retrieval across clinical records.
Flutter Web
Golang
Node.js
RAG
Claude Code
Sample — pending a real client quote“The requirements were traceable end to end before a line of code was written — and the agents built to them. It changed how we scope delivery.”
03 — Top Posts