Case Study: The Trade Desk achieves scalable natural language translation and web page tagging with Databricks

Enabling natural language translation at scale

The Trade Desk, the largest independent demand-side platform, needed a way to include non-English web pages in its contextual targeting and pricing workflow. Half of the pages it analyzed each day were not in English, limiting its ability to tag content accurately and expand reach. To solve this, The Trade Desk partnered with Databricks and used its Data Intelligence Platform to build an NLP model for translating web pages into English.

Databricks helped The Trade Desk move from proof of concept to production in weeks with a scalable NLP solution. The model now processes more than 100 million web pages per day, including half in languages other than English, improving tagging, pricing accuracy, and campaign relevance. This has expanded client reach, strengthened The Trade Desk’s competitive advantage, and created a foundation for future NLP initiatives.

The Trade Desk

Xuefu Wang

Senior Data Scientist