Please submit your answers by emailing them to [email protected] in a new thread. You should link to an external Notion or Google doc. We expect you to spend a few hours on this, producing a few hundred lines of English and code to showcase your talent.

Gumroad Discover is not personalized using prior search/purchase history, nor does it use fancy neural nets to make better recommendations. Explain, via a Notion or Google doc, how you would approach using your background in ML to improve discovery on Gumroad, including on:

• Gumroad Discover, by showing more relevant results
• Product pages, by showing similar products
• Gumroad Library (web + apps), by showing similar products to those they already own
• As part of the Checkout flow, perhaps

Start with what you see as low-hanging fruit within the Gumroad product before moving up to implementing more recent advancements in deep neural nets.

Creating value quickly for our creators is key.

Must-dos include:

• Thoughts on low-hanging fruit that can be handled within our Rails app
• Thoughts on shape of training process for deep neural net
• Demonstration of prior experience solving similar problems in production

A good response shows off your ability to:

• Understand the problem Gumroad solves, and our business model
• Read through API docs