®2026

AR Sephora Shopping with Snap

From June to December 2025, I worked as a design engineer on FastLane, a Snap-sponsored AR startup bringing e-commerce personalization into physical retail. I turned our interaction flows into a working Sephora shopping prototype on Snap Spectacles.

The project began as a hackathon idea and grew into a startup that raised over $30,000 from Snap Inc.

project type

AR Startup | Snap

Team

Isabella Wang

Cody Chen

Edna Ho

Abdi Ambari

Year

2025

Role

Design Engineer

Design for Spectacles

Physical stores offer touch, scent and instant gratification but no personalization. Online stores know what you like, but you can't hold the product. FastLane adds a digital personalization layer to the Sephora aisle, so shoppers get reviews, comparisons and offers without reaching for their phone.

Designed for a busy store aisle, where the product stays in focus and information plays a supporting role.

Glanceable

Keeping product cards short and readable in a narrow field of view

Ensuring readability and safety

during the driving process

Letting shoppers ask for details by voice or gesture while their hands hold products

Hands-free

Anchored

Pinning information to the product in space, so it follows what shoppers look at

System Architecture

Our research at Sephora (field visits, diary studies and shopper interviews) shaped the system. The front end runs in Lens Studio: product recognition, gesture and voice input, card pop-ups, store mapping and the recommender interface. A MongoDB catalog holds the product data, and Gemini Live and the OpenAI Realtime API answer voice questions with live product information.

Figma to Lens Studio

One component system

Our UI was split across designers. I rebuilt their Figma screens as reusable TypeScript components in Lens Studio, so every card shares layout, motion and data binding.

Designing within hardware limits

On the waveguide display, dark tones turn transparent and fine detail disappears. We tuned contrast, type size and card density on-device through repeated builds.

Features

Recognize

Pick up a product and the glasses show its name, price, rating and top reviews.

Compare

Hold up two products and a side-by-side card weighs them against what you care about.

Recommend

Keep browsing and suggestions adapt to the aisle you're in and the products you've looked at.

AR Sephora Shopping with Snap