Engineering Production NLP Systems at T-Mobile with Heather Nolis - #600

This article discusses Heather Nolis's work at T-Mobile, focusing on the engineering aspects of deploying Natural Language Processing (NLP) systems. It highlights their initial project, a real-time deep learning model for customer intent recognition, known as 'blank assist'. The conversation covers the use of supervised learning, challenges in taxonomy development, the trade-offs between model size, infrastructure considerations, and the build-versus-buy decision. The article provides insights into the practical challenges and considerations involved in bringing NLP models into production within a large organization like T-Mobile.

• •T-Mobile deployed a real-time deep learning model for customer intent recognition.
• •The project involved supervised learning and the development of a taxonomy.
• •The article touches upon the considerations of model size, hardware, and the build vs. buy decision.