Launch Announcement: Wayfair Deploys ThirdAI’s UDT to Improve Search Relevance and Reduce Costs

ThirdAI’s Universal Deep Transformers (UDT) AutoML library allows Wayfair to train query classification models on low-cost CPU hardware and serve real-time predictions with sub-millisecond latency

ThirdAI is excited to announce the launch of our collaboration with Wayfair, a leading online retailer for furniture and home goods, to better serve highly relevant search results in response to customer queries.

By leveraging ThirdAI’s Universal Deep Transformers (UDT) toolkit, Wayfair successfully trained a query classification model for search relevance on low-cost CPU hardware and successfully deployed the serialized UDT model in production with sub-millisecond latency. We are especially excited about this collaboration because we were able to achieve a triple win for Wayfair: (1) improving the relevance of search results for shoppers, (2) reducing inference latency, and (3) reducing training infrastructure costs by eliminating the need for costly GPU machines.

For more details on the business importance of product search and the journey of ThirdAI’s engagement with Wayfair, please see the excellent writeup published on the Wayfair Tech Blog.

Use ThirdAI to Transform Your Business

We are still in the early days of our journey to radically democratize AI by enabling businesses and developers to train and deploy state-of-the-art models on ordinary CPU hardware with no feature engineering or parameter tuning. As we have seen in our work with Wayfair, ThirdAI’s software can improve upon existing techniques at a fraction of the cost while seamlessly handling the demand of high-traffic online serving systems. To try out UDT for your business needs, please reach out to us by requesting a trial license for our software and explore our Google Colab demo notebooks that showcase UDT in action.