Smartly Detect Product Defects in Production
Minimising the amount of scrapped material by identifying early which products cannot be fully finished
Challenge
One of the world’s largest automotive manufacturers
Minimising amount of scrapped material during the riveting process helps to reduce costs
Defects could be identified by collecting relevant data from the assembly line in real time
Wanted a solution to empower assembly line workers to identify defects during the riveting process
Built data pipelines to collect the relevant data from the assembly line
Trained a classification Machine Learning model
Designed a dashboard for easy interpretation of model’s outputs
Solution
Value
Empowered workers to identify riveting defects in real time, allowing for the manufacturing process to be adjusted much more efficiently
Contributed to the overall reduction in costs for the manufacturer
Roles
Data Scientist, Machine Learning Engineer
Tools
Sector
Automotive