CONSTRVCT isn’t just a website, it’s a fashion creation machine. I invented a process for making fitted dresses with all-over printed fabric. Users could upload an image and position it over a 3D model of a dress. When a design was ordered, our software exported a print resolution file of the dress pattern with the mapped image, which was inkjet printed on fabric and then cut and sewn into a dress. The distinctive feature of our process was being able to map an image seamlessly across multiple cut pieces of a tailored dress, an exponential step above what the standard custom t-shirt company does.

From a business strategy point of view, on-demand production of apparel is a paradigm-shifting idea. Fashion is a consumer category marked with incredibly high frequency of design cycles. Inventory quickly loses value due to the short seasonality of the goods. Being able to sell a wide range of designs without holding inventory is a great advantage.

With outstanding traction in the first 3 months, we were accepted into the Techstars Boston accelerator class of 2013. We were named one of the most exciting startups of this class by Business Insider and Xconomy.

“*IT’S AN EXPLOSION in the Tumblr-viral street-new-aesthetic fab lab.”

–Bruce Sterling on

CONSTRVCT was featured on Creators Project, The New York Times, Wired, Huffington Post, Slate, Mashable, as well as many other media channels. Our initial press coverage gave us a boost to achieve organic customer acquisition without paid marketing, which was sustained via social channels.

I designed and coded the website in Ruby on Rails and Three.js. I also designed the logo, took photos, designed all the clothing as well as the 3D models, and wrote core image processing scripts. I also managed main operations and customer support, because startup.

Product Iterations

The first prototype of CONSTRVCT was built over a weekend at the Gilt Fashion Hackathon, which my team won.

I built the second prototype for a Kickstarter campaign, which was successfully funded at $11,000. As well as validating consumer interest in the idea, the funding covered much trial and error on clothing production.

The first release of CONSTRVCT.COM had most of the current features, as well as custom sizing. Yes we were crazy to do a custom fit sheath dress. But it did work. Even if the fit was not 100% perfect, there were happy customers who were pleased to be able to get a unique dress in their size at all. I later removed custom sizing because we did not have the operations to support it, but the brief time in which we tried this ambitious idea was informative on customer needs.

In further iterations I gradually improved the UI, re-organizing the main design creation page for clarity and adding things like a separate products list page to meet frequent customer questions. From reading many customer emails as well as conducting a small user survey, I found that there were two main user groups present on the site.

  1. Consumers– These were women who saw the site as a customization brand. They purchased clothing either for themselves or for gifts. These were our best customers, and they made repeat purchases to try the different styles. There were 2 groups within this group: younger tech savvy women who often created designs using images from science or nerd culture, and older women who often where photographers or artists as a hobby or as a second career.
  2. Professionals– Many users were people who wanted to use the platform as a way to create and sell their own designs. (While many were not likely professional designers, the term is chosen because their intent was for business.) Although we never advertised that we were a marketplace, we received thousands of registered users who assumed that a marketplace existed and that they could sell their designs.

Although there is some overlap between the two user groups, the large difference in primary goals meant that likely they would have been better served on two separate sites. If the product progressed further, I would have concentrated on building out the consumer-focused destination first. Because even though there was much demand for a design creation marketplace, you need consumers who want to buy designs in order for a marketplace to succeed.

There were many more ideas for Constrvct, but the biggest hindrance was manufacturing. Realizing that my skills reside on the digital side, I am currently exploring licensing the software platform to other manufacturers. Sign-ups on the site are disabled, but at last count there were about 12,000 users and over 35,000 designs.

A possible next iteration of Constrvct would focus on consumer-friendly features like being able to choose variations in neckline and length, while curating artwork for the prints.


In December 2014, the Smithsonian Cooper Hewitt National Design Museum selected CONSTRVCT to be shown as part of its inaugural exhibition “Beautiful Users”. I created a customized interface for display using artworks from the museum collection and formatting the UI to work on a small tablet. The curator, Ellen Lupton, also wore a custom CONSTRVCT dress for the opening reception.

Felicia Day wore a CONSTRVCT dress for the cover photo of her book. This was a scenario perfectly exemplary of the internet. An LA stylist browsing the site selected a dress design featuring a photo of an op-amp circuit constrvcted by a microbiology PhD student at UC Davis. The dress was then rush printed and sewn in 6 days and overnight shipped to LA for the photo shoot with the geek culture actress.

A Boston paper did an awesome photoshoot with one of our dresses and an MIT robot.