Joshua Pendragon

I'm an AI/context engineer with a quality engineering backbone. I build LLM-enabled workflows, developer tools, evaluation loops, and agentic systems. The point: help engineering teams move faster without losing reliability.

My path started in support, operations, and SDET work. That background shapes how I think about AI systems. Making a workflow understandable, testable, repeatable, and safe for real teams is the hard part. (Demos are easy. Shipping is hard.) I've modernized automation systems, built CI/CD quality gates, reduced flaky feedback loops, and mentored engineers through better tooling. Started in the US Navy as a Cryptologic Technician, which is where the attention to detail comes from.

Recently my work has centered on AI-assisted engineering and context-heavy developer workflows. At SecurityScorecard I built AI solutions for technical document translation to Japanese. I also evaluated and built coding agent setups and best practice docs, and chaired the AI Adoption Culture Committee. In my own practice I use Claude Code and Codex with custom agent workflows: multi-agent orchestration, knowledge distillation, prompt/context systems, and code-quality tooling.

I'm probably best suited for roles where "AI engineering" means building the systems around the model. Context design, tool boundaries, evals, retrieval workflows, agent instructions, developer experience, CI feedback, observability, human-in-the-loop controls. My QA/SDET background fits because AI systems need the same discipline as production software: clear inputs, measurable outputs, regression checks, and fast debugging.

When I'm not coding, I'm usually on Discord running WoW or FFXIV, working through my Steam backlog, or practicing a tune from Stronghold on the mandolin.

Proud supporter of the Software Craftsmanship Manifesto. A good network is the backbone of a successful career, so reach out and say hello.