Client: A technology startup creating intelligent applications for Industry 4.0
Industry: Manufacturing, Industry 4.0
Scope: Algorithm design and end-to-end machine learning development
Key technologies: Google Cloud Vertex AI, Google Cloud Storage, Google BigQuery
Challenge
The core challenge was to create an advanced algorithm capable of optimizing the production process. It had to:
- Analyze key variables to recommend optimal machine configurations.
- Accelerate machine changeovers and assist operators in real time.
- Demonstrate that the production process could be supported dynamically, without delay.
Facing comparable requirements in your own environment?
Solution
Neontri delivered and tested a cloud-based machine-learning algorithm that uses production data to optimize machine settings.
The system evaluates inputs – such as temperature, humidity, and raw material type – against clearly defined goals and key problem characteristics. Once cleaned and analyzed, this information is fed into the model, which either flags data gaps or recommends the most suitable machine configurations.
Built for dynamic manufacturing environments, it combines several key capabilities:
- Predictive analytics: Uses real-time data to determine the right machine settings for operators.
- Process optimization: Directly accelerates machine changeovers, reducing downtime and increasing throughput.
- Automated MLOps: Automates model testing, training, and optimization through Vertex AI.
- Real-time data processing: Leverages BigQuery to handle large, continuously growing datasets as new records arrive.
Delivery approach
We collaborated closely and transparently with the client throughout the project. As requirements evolved, our team adapted quickly and took a proactive approach to resolving the technical challenges of real-time industrial data analysis.
Results
The project demonstrated the tangible value of applying machine learning to industrial challenges, delivering clear, measurable outcomes:
- 75% accuracy in predicting the correct machine settings
- Faster model training and optimization with VertexAI compared to on-premise setup
- Lower long-term costs for keeping the infrastructure running
- A flexible, cloud-based approach to production-line configuration
Want to see what real-time ML can do for your production line?
Team up with specialists who have built scalable, cloud-native optimization systems for industrial manufacturers.
Our process for strategic partnership
- 01
Discovery
Business goals, requirements, and constraints are reviewed to shape the project’s delivery approach.
- 02
Design
A scalable, secure architecture is defined to support both immediate needs and long-term growth.
- 03
Development
The solution is engineered using modern practices, with quality maintained at every stage.
- 04
Testing
Rigorous testing validates performance, security, and reliability before release.
- 05
Deployment
The product is rolled out through a controlled process that minimizes disruption for users.
