I make complex tools and ideas feel clear, useful and human.

I'm a product and UX designer who makes complex data and AI products feel simple. At Infodat I lead design for data tools end to end, from user research and information architecture through to shipped, high-fidelity UI.
My work spans product, growth and marketing: interface design, design systems, landing pages, marketing funnels and brand.
An AI-powered ETL and data intelligence platform that lets people load, transform and question their data in plain language.

Most data work assumes technical skill: writing SQL, wiring pipelines, knowing how the data is shaped. That locks out the analysts, operations leads and business users who actually hold the questions. Morphus lets them load, transform and explore data without ever querying a database.
Ask questions of your data and get answers and charts back, each one traceable to its source.
A node-based flow for loading and transforming data, with real-time feedback so errors show up while you build.
A tutorial-style onboarding where new users build and run their first pipeline. Experienced users can skip it.
The component library, type and color tokens, and layout patterns the product is built on.




A brand identity and design system for Akeepo, an early-stage technology and consulting company.

A young company needed a distinctive identity and a system that would hold up as it grew.
Inspired by 1960s and 70s computing and Brutalist architecture: blues, greens, oranges and purples, bold geometric shapes and grids, and vintage terminals.
A memorable look that presents Akeepo as a tech company that values both where computing came from and where it's going.





A research app for NBA player props. Its main workspace, NBA Bet HQ, puts five seasons of game logs, lines from the major sportsbooks and the live box score on one screen, so a bettor can check a prop in minutes instead of going on a hunch.

Most people bet props on a feeling. A player hit his last three, the line looks low, the group chat is all over it. That sometimes works for a night but rarely over a season. The numbers that would settle it are spread across box scores, injury reports and ten sportsbooks, and nobody with a job has an hour before tip-off to pull them together.
People who would rather check first: college students stretching a small bankroll, and people with jobs who have five minutes, not fifty. They know the players but not the math, so the app has to do the math and show its work in plain language.
Every pick traces back to hit rates, a projection and line movement. Slip details list each factor that set the stake, down to the dollar.
Green cleared the line, red fell short, and the dashed line is tonight's number. Every chart, chip and tile follows that rule.
Edges take a market haircut, stakes use quarter Kelly, and picks that look too good get marked down. ROOKIE, THIN DATA and OUT tags flag weak numbers.
Per-slip and per-day caps, 21+ and 1-800-GAMBLER on every screen, and copy that never promises a win.








BetBtr is invite-only while it's being finished, with a personal pass key for each member and a free preview of tonight's board for everyone else. On a backtest of 500 of the day's lines against each player's last ten games, model picks hit 64.2% against a 57.8% break-even. That is a check on recent games, not a promise, and the app says so.