AI-Driven Automation for Quality Control in Dairy Production Systems Intern
AI-Driven Automation for Quality Control in Dairy Production Systems Intern
- Do you want to support us in defining the next generation of Milk Foaming Appliances for the Horeca market. FrieslandCampina Lattiz makes it very easy for all professionals to make top quality cappuccino's or any milk containing coffee recipe.
- We offer trainee positions within the Development&Technology Professional Systems team.
What we ask
- You are an ambitious HBO/University student with a background in electronics, software, or mechatronics.
- Availability for a period of about 6 months
- Skills in Python, TensorFlow, Keras, or similar machine learning libraries.
- Interest in working with image recognition and Artificial Intelligence
- Experience with real-time systems and streaming data is desired.
What we offer
We offer an inspirational work environment in the R&D team of Lattiz. The team is located in Wageningen in the FrieslandCampina R&D facilities on the Dairy Campus.
Vacancy description
Assignment Description: Automation of Foam Quality Control for Milk using Convolutional Neural Network (CNN)
Background:
Currently, the quality control of foam made from milk is performed by baseline testers in the lab (experts). This process is prone to human error, variability due to human interaction, and is not optimal for consistency and scalability. To improve these processes, we aim to automate quality control using AI, specifically a convolutional neural network (CNN), that can reliably assess and report foam quality in real-time.
Objective:
Develop an AI system based on a convolutional neural network (CNN) that automates the quality control of foam in milk products. This system should be able to analyze multiple quality parameters of foam in real-time and compare them to an existing quality standard. The goal is to reduce the error-proneness of the current manual process, minimize variability, and improve the efficiency of quality control.
⏳ Duration and Planning:
- Time Limit: 6 months.
- Phase 1 (0-2 months): Research the existing setup, collect training data, and evaluate different CNN architectures.
- Phase 2 (2-4 months): Develop and train the CNN model. Iteratively improve the model's accuracy and ability to assess foam quality.
- Phase 3 (4-6 months): Implement real-time processing, integrate the system into the existing production environment, and finalize the system with a user-friendly interface.
Deliverables:
- Final report with documentation on the developed AI system, optimized setup, and conclusions on the effectiveness of the automation.
- A working prototype of the AI system with real-time foam quality analysis.
Team Details
For thousands of people every day, we are more than just a dairy company. To our farmers, our employees, the communities we serve, the businesses we work with and the people to whom we bring happiness, FrieslandCampina means something more. For them it's not just about what we do, but who we are.
We value talented people from any background who want to contribute to something bigger than themselves. We encourage all of our employees to make decisions that benefit our entire company. At FrieslandCampina we own our own career and act accordingly. We trust you to make a difference in your job and influence the bigger picture. Working at FrieslandCampina means you are contributing to a better world.
The team is consisting of around 15 people on the FrieslandCampina location at the Dairy Campus in Wageningen.
