I design products, build them, and sell them.

JLC sells in drops. Each launch sends a spike of mostly mobile visitors who decide fast. Any friction on product pages, in the cart or at checkout costs sales that don't come back after the drop ends. The off-the-shelf theme also made JLC look like every other Shopify store, which hurts a brand that positions itself as luxury streetwear.

Goal: get shoppers from a social post to a completed checkout with as little friction as possible, in a look that's unmistakably JLC.

1. Find where shoppers drop offI used Shopify session and funnel data to see where visitors left: product page, cart, or checkout. I also looked at device split and referral source.
2. Fix friction on mobile firstI debugged and fixed a cart error (HTTP 422) caused by a deleted shipping zone, which was silently blocking US orders. I fixed inventory display bugs across several theme files and made the checkout button in the mobile cart drawer always visible.
3. Simplify navigation and product pagesI cut the landing header down to search, account and cart. I rebuilt product pages so size, stock and price can be read in one glance.
4. Build a custom theme and brand systemI'm coding a fully custom theme from the Rise base, with a dark arcade look, one pink accent for every hover and call to action, a Snake-game landing page that spells out the brand name, and a lookbook gallery page.
5. Plan launches as product rolloutsI plan each drop as a staged rollout. For the Goyadongo bag: Monday, a teaser; Wednesday, 20 numbered units revealed; Friday, the site goes live with campaign photos. I measure sell-through and use the results to decide the next drop.

• $47.9K in total sales across 494 orders and 370 customers, with an average order of about $89.
• Conversion rate reached 2.4% in July 2026, above the typical e-commerce average of roughly 1.5 to 2%.
• The July 2026 drop did $18.3K and 124 orders in one month. October's drops added $8.5K.
• About 1 in 10 customers have come back to buy again.

Small service businesses miss calls and then enter leads by hand. Orbyn wanted an AI agent that answers the phone, qualifies the caller and books the appointment, with every call ending up as a clean record in a CRM. Each Orbyn client needed their own data walled off from the others, and different staff needed different levels of access.

• The UX architecture and all front-end screens in React: contact profiles, dashboards, search and filtering, and admin workflows.
• Tenant switching and role-based views. Users see only what their role allows, and every screen makes clear which client's data is showing.
• Call summaries next to each contact, so staff can read what the AI handled without replaying the call.
• Followed established SaaS design patterns and accessibility basics (contrast, keyboard navigation, clear labels) so non-technical staff could use it on day one.
• Helped plan the data model for secure multi-tenant storage with FastAPI and SQLAlchemy.

We delivered a working proof of concept that is now part of Orbyn AI's production site. Orbyn uses it to show prospective clients what their automation can do.

1. Concept & referencesStart from a cultural reference and a silhouette, then define the colorways and the one detail that makes it ours.
2. Tech packWrite factory-ready specs: materials, hardware finish, lining, trim, label and branding placement.
3. Sampling with manufacturersOrder prototype samples, review them, and go back and forth with the factory until it's right.
4. Campaign productionWrite shot lists, cast and direct models, and coordinate photographers and video shoots.
5. Launch & measurePlan the rollout calendar, run the Meta ads myself, then review sell-through to plan the next drop.