WO: Internship Data Science - Classification of printhead failure reason

Are you a WO data science student who is looking for a graduation assignment? Are you ready to delve into the data to find why a printhead was replaced and automate the decision? We are looking for you!

Your assignment
Canon Production Printing (CPP) develops and produces a variety of commercial size printers for the business to business market.  Our  inkjet printers are designed for high volume printing through a combination of high speed and low operational costs, while retaining high quality and flexibility.
The printhead is the droplet forming unit, the heart of our printers. Failing of the printhead directly impacts print quality. This has large consequences for our customers and is therefore undesirable. On top of that, replacement of a broken printhead is a very costly event. 

The project
When a printhead fails, the logging data is an important source of information to determine the failure mechanism. Sensor data, error data, performance of the individual drop forming nozzles or usage data might point to a certain failure mechanism. But in some cases the logging shows signs of multiple failure mechanisms. 
Goal of this project is to come up with a method to determine, from the different logging  data tables, the most probable failure mechanism. 

Challenges for you are:
-    Develop a method to classify the failure mechanism of a replaced printhead automatically
-    Can you come with a way to deal with yet unknown failure mechanisms?
-    Apply the method to printheads that have not been replaced yet to already give a warning about upcoming failure 
-    Can you deal with large amounts of data. Besides error codes and usage data which are logged with quite low frequency (hours) , sensor data (including status of hundreds of individual nozzles) is typically logged every few seconds.
-    Deal with relatively small amounts of failures. Although data is available for thousands of printheads, no more than 20 or 30 failures (or less) may be found per failure mechanisms. 

Your profile

•    You are currently studying Data Science or Data Analytics and you are looking for your MSc internship or graduation assignment at a big international company
•    You have a critical attitude and you are curious to understand your own observations.   
•    You have at least basic Python and SQL knowledge 
•    You are available for 5 to 8 months for a minimum of 4 days per week

What’s in it for you?
•    A challenging assignment with skilled coaching
•    Internship/ Graduation compensation of €500,- per month
•    Travel cost compensation if you don’t have an ‘OV-weekcard’
•    The possibility to network with professionals inside and outside your field of expertise, thanks to our diversity of disciplines which you will work with
•    With our company values collaboration, innovation and ownership, we strive to bring out the best in each other, expand boundaries and feel empowered to take ownership of our work.

Interested?
Are you interested in this assignment? Please click on the button 'apply now' where you can upload your resume and motivation letter. 
If you would like to receive more information concerning this assignment, please contact Evelyne Janssen, tel. +31 (0)77 359 2694. If you have any questions about the internship in general, please contact Milou Geelen, + 31 6 25653192.

About Canon Production Printing
Canon Production Printing develops and manufactures high-tech printing products and workflow software for the commercial printing market and is part of Canon, a global provider of imaging technologies and services. Canon Production Printing operates on three continents, with approximately 3300 employees (end 2022) and has its global headquarters in Venlo, the Netherlands.

Further information about Canon Production Printing is available at cpp.canon.