# Closed Loop

> AI engineering, evidenced. Long-form technical articles on agentic systems,
> Claude Code, AI infrastructure and AI security by Damian, a senior AI engineer in Sheffield, UK.
> Every article documents a system that ran in production — architecture, hard gates,
> and the failures that shaped the rules.

Damian is available for remote freelance AI engineering work: agentic system
design and build, Claude Code enablement for teams, AI infrastructure consulting,
and AI implementation security audits — security review of AI-generated code and
of the agentic layer itself (prompt injection, tool permissions, data egress).
Contact: hello@closed-loop.dev — see https://closed-loop.dev/services/

## Articles

- [Graph Engineering: A Dev Workflow You Can Drive From a Browser](https://closed-loop.dev/blog/graph-engineered-workflows/): Model an agentic engineering workflow as an explicit graph — states, guarded edges, a dependency DAG, and questions as gates — then give the human a web console that projects it.
- [Folders Are States: Building an Autonomous Dev Loop with Claude Code](https://closed-loop.dev/blog/folders-are-states/): Turn an LLM coding assistant from a chat window you babysit into a background co-worker that clears your backlog and proves its work with screenshots.

## Pages

- [About](https://closed-loop.dev/about/): who Damian is and what this site is
- [Services](https://closed-loop.dev/services/): freelance AI engineering offers — agentic build, Claude Code enablement, AI infrastructure, AI implementation security audits
- [How we work](https://closed-loop.dev/how-we-work/): how engagements are diagnosed (constraint), measured (one KPI) and priced (10–20% of first-year value), with a calculator
- [RSS feed](https://closed-loop.dev/rss.xml)

## Full content

- [llms-full.txt](https://closed-loop.dev/llms-full.txt): full text of every article in markdown