I'm Patrick King. I've spent about twenty years in institutional capital markets, and I build research and decision systems in the evenings — some for markets, some for running a household.
Most of what I make is documented publicly while it's still unfinished. That's deliberate. I write a weekly build log about what I'm building with AI and agent systems, including the parts that don't work, and I publish market research where every count carries the level that would prove it wrong.
What I'm building
- King Nuggets — site live, letters pending Independent market research. Cycle counts published with their invalidation levels attached, so a wrong call stays visible as a wrong call.
- Build Log — publishing weekly Every issue in one place. What I'm learning, what I'm building, and what broke this week. Published as PK Learns.
- King Nuggets Lab — in development The workspace behind the research: methodology, models, and point-in-time archives.
- Projects — live What I've built and what state each thing is currently in. A live look inside my head.
- King OS — internal The automation and agents underneath everything else. Not published, but it's the reason the rest runs.
Now
Updated August 2026
Writing the build log every week, which matters more than any single issue of it does. King Nuggets is where the market research lives — the site is up and the method is written down. The letters stopped when work got heavy; getting them going again is the next thing. And figuring out what to do with the CMT charter I earned this summer.
How I work
I work on messy operational and research problems where data, automation and human judgment meet.
Most of my projects begin with something that sounds reasonable but is still vague. I try to turn it into explicit rules, evidence and a decision: what is true, what changed, what failed, what needs attention, and what should happen next.
Problem → Assumption → Method → Artifact → Evidence → Failure / surprise → Decision
A project is useful when it changes what happens next. The artifact is proof of the work, not the point of the work. Three worked examples →
About
Twenty years in institutional capital markets, most of it close to equity compensation. I hold the CMT charter. Outside the day job I build things and write about building them — market research grounded in cycle work, agent systems that do real jobs, and the infrastructure holding both together.
I'm not selling anything here. If something I've written is useful, that's the point. Bay Miles Group LLC is the operating entity behind the work.
Contact
Email me at patrick@workbypk.com, or find me on X at @pk_learns and on LinkedIn.