From billions of stored photos to real-time product recommendations, Shutterfly is harnessing machine learning at scale to deliver faster, smarter, and more personal engagement across every touchpoint.

Customers come to Shutterfly to create personalized products and gifts that are as unique as they are. And, we believe that our engagement with customers should be personalized and unique too. So whether customers log in on our site, pop open the Shutterfly app, or receive a personalized email from Shutterfly, they’ll find that we’re curating photos they’ll love and matching those photos with products that resonate.

Ever wonder how we make that happen? It starts with the billions of customer photos stored on our platform. Using advanced analytics and smart models, we analyze image types and customer behavior—always in aggregate—to build tools that help people easily curate their memories and create beautiful, personalized products.

In this technical blog post, Databricks breaks down how Shutterfly utilized an open-source framework on the Databricks platform to scale machine learning for personalized ecommerce, achieving:

• 20x faster model training,
• Accelerated model development throughput more than 3x, and
• Real-time personalization enablement,
  all which is driving ease of use and real customer engagement.

Check out the full article on the Databricks website here: Scaling Machine Learning with Ray and Databricks.