I'm building Quillmark, a small operator-led studio for grounded AI systems: tools that stay close to source material, make judgment visible, and leave room for human review.
Most of my work sits around:
- AI workflows shaped by real policies, examples, transcripts, and approved language
- review tools for content where accuracy and trust matter
- source-backed drafting and analysis
- commercial services for narrow, high-consequence workflows
- open source projects that make practical AI work easier to build and inspect
Right now I'm especially interested in agent workflows, grounded content systems, and software that turns repeated judgment work into something more reliable.
You can find my open source work at Quillmark Open Source.