Selected work

My recent work has focused on AI products for complex knowledge work inside Intuit.

I work on opportunities from the beginning: identifying and prioritizing problems, doing discovery, and developing a product direction. That continues through technical solution development with engineering, ML, data science, and design, then launch, adoption, iteration, and measuring impact.

The work has included internal agent orchestration, employee knowledge, forecasting, and workflow agents. Finance and compliance are part of that context. The product questions often depend on understanding both the work itself and what the model can reliably do.

I’m particularly interested in the connection between discovery and evaluation: understanding what a user needs, deciding what the team should test and optimize for, and defining how we’ll know whether the system is solving the problem we chose to prioritize.

A mobile commerce platform for brick-and-mortar secondhand retailers.

We brought $5B worth of commercial goods online.

My work included product direction across iOS, Android, and a web plugin, along with recommendation systems, retailer onboarding, and launch strategy. As the business developed, that also included a shift toward B2B SaaS.

On the founder side, I worked on retailer adoption and partnerships, including Goodwill Industries and the California Pawnbrokers Association, and helped support the company’s fundraise.

I’m building Bill Screener to help organizations keep up with California legislation.

There’s a lot to follow: new bills, amendments, and status changes—sometimes more than 1,000 in a single day. Which ones matter depends on the client or organization. Bill Screener uses client profiles to screen legislation for relevance and brings likely matches into a queue for review.

How the product works

For lobbyists and government affairs consultants managing multiple clients, Bill Screener turns the daily legislative feed into a client-by-client review queue. They can confirm relevant matches, choose which bills to track, follow later amendments and status changes, and have Bill Screener draft client-ready email updates from the bills they select. In-house policy teams can use the same workflow to monitor legislation against their organization’s priorities.

The client profile and review step are central to the product. Relevance is specific to the client, and the person reviewing the results can decide what to follow.

Technical context Next.js and TypeScript, OpenAI for screening, OpenStates for legislative data, and Postgres with Prisma. Deployed with Docker and Caddy.

A marketplace for brand IP and license management.

I worked on brand onboarding, license applications, royalty tracking, and application management. The product involved a licensing process developed with legal experts, connecting the requirements of an agreement with the experience of applying for and managing a license.

A business combining proprietary software with financial education.

I led hiring and operations for a cross-functional team and built revenue streams around education on stocks and cryptocurrencies. The business was later sold.