Quality Control in
Build-to-Order Manufacturing
We are excited to talk about our partnership with PACCAR Inc. We have been working with their Peterbilt division to further improve their excellent quality control processes, using our computer vision technology.

This partnership has been a great learning experience for us to understand "build-to-order" manufacturing. Traditional machine vision solutions focus on high-volume production where a fixed product, such as an iPhone, gets assembled in very large quantities. In contrast, PACCAR's subsidiaries offer product portfolios comprising 10s of 1000s of product variations, while maintaining high throughput and consistent quality.

A key challenge for automated quality control in such mass-customized production is the need for 1000s of "training examples" for every product variation coming down the line. We reduce this need by defining compositional process models where machine learning algorithms are trained on sub-assemblies and combined together for individual build orders extracted from the Manufacturing Execution System (MES).

We believe that PACCAR's production model will be followed by an ever increasing number of manufacturers as customers' demand for customization keeps growing across discrete manufacturing verticals. Correspondingly, in this paradigm robots have a relatively smaller role to play as they cannot match the adaptability of human workers. Our focus on helping manufacturing associates avoid assembly mistakes, receive better training, and obtain prescriptive analytics to automate Kaizen events precisely targets this paradigm.

Watch this space for several more customer profiles that we will be sharing soon!

Made on