Lorem ipsum dolor sit amet, consectetur adipiscing elit. Sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat.
Staff product designer at Dropbox, leading design for AI—with a focus on Dash as the place work context turns into useful conversation.
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat.
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat.
Live commerce UX for collectors: breaks, auctions, and community moments on mobile.
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat.
As a senior product & design manager at Late Checkout, I worked with the team at Epic Games to develop an RSVP system for Fortnite.
>10m
Estimated viewers of the Fortnite Rift Tour virtual concert
With over 650 million players, Fortnite has grown immensely since its release in 2017, cementing itself in pop culture.
A pivotal moment in the game's trajectory was when it began hosting concerts featuring artists like Marshmello, Travis Scott, and Ariana Grande.
This opened Fortnite up to new audiences and required new UX to facilitate virtual events.
With a broad brief about encouraging gen-z to indicate their intent to attend Fortnite events, I first concepted and pitched a comprehensive "intend to play Fortnite" system.
This direction was well received by the Epic team but soon after, due to time constraints, we set the scope of the project to event-specific UX.
Rift Tour website wireframes
Wireframes defined the product's key flows on mobile.
Rift Tour visual identity
We pushed the identity of the Rift Tour via the product.
Live Rift Tour tab
The Epic team also implemented our work as a bespoke tab in Fortnite.
The event was attended by millions of people and was reported on by numerous tech & gaming publications.
The intent-to-attend tab was heavily trafficked and well-received by the Fortnite community.
With local AI being on-device and free to run repeatedly, I set out to build an 'always on' AI writing assistant but pivoted after uncovering the real bottleneck in my writing: formulating a crisp point of view.
Holm V3: The guided draft
Having explored always-on suggestions and local model notes-to-draft, I reevaluated the experience and realized that the bottleneck in my writing practice was earlier on in the process.
I had plenty to say but needed help clearly articulating my thesis, supporting arguments, and mapping out the structure of my writing. Holm V3 uses frontier models to prompt you to refine your thinking.
This is by far the most useful version of the product yet. I use it multiple times a week.
It turns out that sometimes the models should be prompting us.
Always-on suggestions
Holm V1: Always-on Writing Suggestions
Holm V1 used local models like variants of Gemma 4 and Qwen 3 to provide in-line provocations based on what you were writing.
It worked well for cases like journalling, where the models would ask you questions or prompt you to say more about your feelings, but didn't work as well for the type of editorial writing I was interested in.
Bump suggestions
Local models surfaced inline provocations as you wrote—questions and nudges that showed up directly in the note. Journal-style drafts responded especially well; this clip is the bump-suggestions interaction in motion.
Holm page
The home page uses local models to generate titles for notes, with a few adaptive rules & views depending on note length.
Draft from notes
Guided draft in motion
Screen capture of the guided draft flow—prompting and structure surfaced as you shape the piece.
Guided draft
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat.
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat.
Lorem ipsum dolor sit amet
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat.
Consectetur adipiscing elit
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat.
Sed do eiusmod tempor incididunt
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat.
Signals for showing up—to friends and feeds—without the doom-scroll.
With agents owning the structure of web design, it can be difficult to understand what's going on, and ensure that elements align, and relate to each other in the ways you want. I built Redline to replace using the native screenshot tool as a makeshift ruler, and for quicker access to a clearer DOM.
Dense dashboards
When cards, rails, and chrome stack up, alignment stops being obvious. Redline makes it easy to check spacing and relationships without guessing from a flat screenshot.
Live layout, clear overlays
Measurements sit on top of the real page so you can verify gutters, baselines, and component edges against what users actually see—not a frozen crop.
DOM-aware inspection
Pair rulers with a clearer read on structure so you can relate what’s on screen to the underlying tree—and adjust with confidence.
Autonomous process execution for large language model function buttons
LLM action buttons in the browser that start workflows across your open tabs from the page you’re on.
Graphical user interface for generating large language model function buttons
A browser UI for defining scope and turning it into custom LLM function buttons.
Generating coaching insights using large language models to improve pulse status
Pairing structured work data with an LLM to surface timely coaching prompts.
My first love was making things.
I went to a public fine arts high school and learned to make drawings, ceramics, etchings, photographs, and paintings. Then OCAD University where I learned to make furniture, toys for kids with special needs, and websites. Now, I make AI-powered apps that lighten your load at work, and fun, experimental ones to use throughout the day.
My second love was technology.
I believe in technology as an evolutionary force, as described by Kevin Kelly. I believe that it's in our nature to build tools that improve circumstances for ourselves and others.
Design is how I combine my first and second love.
I believe in design as a way of expressing love to the person using your product, as described by Jony Ive. The best way to demonstrate care for someone is to show up for them, solve a problem for them.
I wake up excited every day to make something. I'm grateful for the journey so far, and for the support along the way.
Jobs
Dropbox
Ocoder
Late Checkout
