AI work
I build with AI, not only advise on it. The clearest example is a research platform I built for my own investing, shown here for the method, not the market calls.
The platform is a personal project. The screens are unbranded single-day snapshots from October 2026, the same ones on the deal sheet, shown for the method and not as recommendations. The last band draws on in-house work at PIMCO.
01 · Personal project, 2026
A research platform run by AI agents
The situation
Following digital-asset markets well means reading dozens of sources every day, from prices and derivatives to fund flows, liquidity and macro data. Nobody reads all of it every morning.
What I did
I architected and built a research platform that orchestrates 25+ AI agents across 60+ market and macro data sources. It covers liquidation clusters, funding and open interest, options positioning, stablecoin liquidity, ETF flows, and a cycle and scenario framework built on price history since 2014.
What came of it
The whole daily read comes together in one place, every day.
02 · Personal project, 2026
Dashboards that say where things stand
The situation
Raw data is not a read. The useful view is a handful of numbers that say where things stand and how far they tend to move.
What I did
I architected and built the research platform these dashboards run on. They show cycle position, futures liquidations, stablecoin supply, and how often Bitcoin has moved 3, 5 or 10 percent within 3 to 14 days.
What came of it
One screen that shows where the market stands each day.
03 · Personal project, 2026
Plain-English reports from the same data
The situation
People decide from a page, not from a model.
What I did
I designed technical-analysis tooling that ranks the leading crypto assets daily, and a one-page report that reads trend and exhaustion counts across daily, weekly and monthly timeframes in plain English, with every key price level.
What came of it
A one-page read for each asset that anyone can follow.
What this means for an AI strategy review
The same questions come up inside a large company. Where does AI create value, what has to be true first, and what does it cost to keep running? An AI strategy review answers those in writing, sized and argued.
The habit is older than the tools. At PIMCO I led a cross-functional team that built an internal application giving real-time access to client account balances and transactions, which turned quarterly reporting into something people could check any time.
Related
How engagements work →Fixed-scope work →