Senior / Staff Product Designer

Most products are built around the wrong thing.

I start one level under the screens: what is this built around, and is that the work? Ten years of complex B2B SaaS, Senior IC through Director. I design AI behavior as a system, then ship the production code. Embedded with engineering, or solo when that's the shape of the team.

Recent work

Contractor CRM Sole designer and developer across three products — UX through the database. The incumbent’s nouns turned out to be states.
Grimoire AI system design, end to end. The coach proposes; code stays the system of record. Fifteen days to production.
Ascend Behavioral product invention. Five constraints enforced in code, not copy. Shipped to the App Store.
Carlyle Lead UX inside an enterprise product team. A schema-first platform rebuilt around three analyst roles.
Cereba The call to rebuild, not retrofit. Conversational onboarding that replaced the blank prompt.

See case studies → How I work →

Selected Work

Five projects. Each one earned the page.

Contractor CRM · Design partner · 2026

A Lead was never an entity. It was a job that hadn’t started yet — and finding that out deleted three tables, two route trees and a whole section of the product.

A CRM, estimating, and field platform for roofing and exteriors, built to replace the incumbent every contractor already knows — and built on its vocabulary, which turned out to be the problem. Sole designer and developer across three products on one design system — UX, front end, backend, and schema — with the density rules written where they can’t silently rot.

Scope: Three products, one system
Role: Sole designer + developer, with the founder
Status: End-to-end testing

Read the case study →

The scheduler board: crews as rows in trade swimlanes, with coloured bars spanning days across a week and an unscheduled tray below

3 products

Grimoire · Solo design + build · 2026

A production Magic: The Gathering deck builder with an AI coach. Strategy, design, code, and go-to-market from one seat — shipped in 15 days, live in production.

Stands on parity with Moxfield and Archidekt and wins on one wedge: an agentic AI Deck Coach that builds a legal Commander deck from the cards you already own. 66,761 lines of TypeScript and 406 commits at ship; 1,043 commits and 314 test files today, with the distribution mechanics in the codebase from week one.

Build: 6 weeks solo
Scope: Solo end-to-end + GTM
Stack: Next.js + Claude

Read the case study →

Grimoire's deck workspace: an Atraxa Commander deck with live prices and a build plan, beside the Deck Coach rail mid-analysis

6 weeks

Ascend · Solo design + build · 2026

Solo designer-developer. Shipped an AI app to the App Store. The product catches the moment a user freezes at their to-do list and returns one next move.

An AI-native app that catches the moment someone freezes at their to-do list and quietly returns one micro-step. Publicly on the App Store; built end-to-end as solo designer-developer, and designed by subtraction — no streaks, no lists, no nagging.

Shipped: App Store, Jun 2026
Scope: Solo end-to-end
Stack: Next.js + Claude Code

Read the case study →

Ascend's landing page: the headline 'You know what to do. You still can't start.' beside a phone showing a completed Clarity Loop

1 action

Carlyle · Lead UX · 2025–2026

Got three analyst tiers out of Excel and into the platform. Role-based workflows from entry analyst to leadership review.

The schema-first MVP forced every analyst into the same Excel export loop. The redesign built three distinct workflows — entry, senior, leadership — each tuned to the work that role actually does, with AI-assisted calculations in-platform. The export loop went away.

Role: Lead UX
Scope: Three analyst tiers
Outcome: Excel export loop replaced

Read the case study →

Generalized view of the Carlyle AI investment platform: assumptions, fund inputs, and version history

NDA

Cereba · Principal AI/UX Designer · 2024–2025

Conversational AI that builds itself. Live in 20 minutes, zero prompts.

Solved the blank-prompt problem with a conversational onboarding system (C.O.R.E. Intent). The user answers questions; the AI configures itself. No prompt engineering, no training data, no setup. Just answer and launch.

Time-to-live: 20 minutes
Setup: Zero prompts
Role: Principal AI/UX

Read the case study →

0 Prompts

Selected projects only. Ten years of additional work available on request.

Approach

I design in code now.

Most AI products wait for a prompt. The ones I design notice what's happening and move first.

Research, strategy, user flows, and information architecture first, built as documentation in the repo with AI. Then design in code, with Cursor, Claude Code, and Claude Design in the loop. Prototype, ship, validate, iterate on data. Figma stays in the toolkit when collaboration or design-system work needs it. Real user research stays in the toolkit when the problem needs direct insight. The point is the full loop: discovery to production to validation, all in one place.

Cursor · Claude Code · Claude Design · Antigravity · Figma · Next.js · Tailwind

Read the full workflow →

About

Senior UX through Director, across complex B2B SaaS. Research, strategy, information architecture, and systems design across every role. Most recent year: Lead UX at Carlyle on a confidential AI investment platform for analysts and senior execs, the work that got them out of Excel and into the platform. Now a designer-developer working on agent UX, conversational interfaces, and the design systems that make complex platforms legible. I run discovery, test with users, ship production code, and iterate on data. Embedded with engineering, or solo when that's the shape of the team. North Smithfield, RI. Remote-only, Senior / Staff / Lead IC.

Off the clock: lives in North Smithfield, RI with his wife and two kids. Coaches his son's basketball team. Plays Magic: The Gathering, reads Marvel Comics, still draws by hand.

More about me →

Where to next