SPOT OM: an AI runs the account, a human signs twice
An AI operating system running a real Instagram account. Two human yeses per post, rules locked in tools, and zero fabrications published in six weeks.
Senior Product Designer at J.P. Morgan Chase, New York. Regulated fintech by day, AI trust systems on the side.
An AI operating system running a real Instagram account. Two human yeses per post, rules locked in tools, and zero fabrications published in six weeks.
The cloud platform replacing every International Private Bank system. One stage pattern runs Payment, Lending, Asset transfer, and Deposit.
Three applications merged into one service desktop for 837 client service users, with trust as the first thing on screen.
I design products people can trust.
Trust lives in the structure: who decides, what gets checked, and what can always be undone. Then I keep refining the system so it does more for people and takes the manual work off their hands.
Strict rules, many hands on one ticket. I design clear ownership, controls that cannot be skipped, and adoption across tenured and new users. Fraud risk drops when the secure path is also the easy path.
AI is capable, but it is not honest by default. That gap is design work. Human approval sits at the expensive decisions, safety rules live inside the tools, and the strongest fix is a system where the error cannot happen at all.
Senior product designer at J.P. Morgan Chase, leading Asset transfer, Lending, and the design system in the Global Private Bank. The authentication experience I led raised authentication success 88% in five months and grew high-security adoption 357%. On ACE, my shared workflow pattern cut completion time 14%, on a WCAG 2.2 design system shipped through Storybook.
On my own time I built SPOT OM, an AI that runs a real Instagram account. When it invented a product and marked it verified, I redesigned the system so that cannot happen: two human gates, rules locked in tools. Six weeks live: 51 posts, zero false claims published, nine hours of weekly work down to seven minutes.
What I do underneath both: strategize through complexity, design the risk out, and build for the scale that comes next. On the side I keep learning where AI is going and what it will do to how people experience everything, one step at a time toward making the imagined real for everyone.