Case Study: Bos-Dimex achieves multilingual predictive text classification with Inoxoft

Bos-Dimex processes 2 million websites per day with Inoxoft

The customer, Bos-Dimex, a global AdTech company, needed to analyze complex websites to improve ad targeting accuracy but lacked the capability to process millions of pages daily. They engaged vendor Inoxoft to develop a machine learning solution for predictive text classification.

Inoxoft implemented a multilingual NLP model using Tensorflow that categorizes web content into IAB categories. The solution successfully processes two million websites per day, obtaining inferences in under 10 milliseconds, which allows Bos-Dimex to serve highly targeted ads and successfully sell their data services to large brands.

Bos-Dimex