Selection, testing & implementation of a suitable ML management tool by valantic
Professionalization of ML life cycles
When a company uses multiple machine learning (ML) applications, a tool is needed to coordinate them. valantic developed a tool for a logistics service provider that simplifies internal development processes involving ML applications.

Challenge
The biggest challenge in using ML applications is the manual effort required to manage their lifecycles.
Consulting approach
An ML management tool can help minimize the labor-intensive handling of data. To select the right tool, it takes a systematic analysis of the technical requirements.
Client benefits and solution
Employees of the logistics service provider can use the tool implemented by valantic to manage the lifecycles of their ML applications more efficiently.

The Challenge in Detail:
ML applications can be highly useful for everyday tasks. However, the more you use them, the more they accumulate large amounts of data and lifecycles. If these ML lifecycles are not systematically maintained – for instance with the help of an ML management tool – the best model for a specific use case has to be selected manually. You can imagine the workload this creates when dealing with hundreds of ML applications and thousands of trained models. Therefore, valantic was tasked with selecting, testing, and implementing a tool based on user requirements.

Consulting Approach:
valantic supported the selection and integration of a suitable ML management tool for the company. The following steps were necessary:
- Requirements analysis
In the first step, all technical and general requirements were identified. This helped determine the features and capabilities the tool needed to have. - Architecture definition
After selecting a tool, the next step focused on the system architecture. valantic also assisted with the integration with the existing IT infrastructure. - “Hands-on” testing
The selected tool was tested extensively, using two internal ML applications (use cases). - Final step: implementation
After testing and optimization, the ML management tool was integrated into the company’s internal development processes.

Solution and Client Benefits:
The advantages of an ML management tool are clear: The manual workload involved in managing ML lifecycles is significantly reduced. This allows human resources to be allocated more efficiently elsewhere. In addition, the best models for a specific use case can be identified more quickly thanks to systematic processing.
Your Contact

Laurenz Kirchner
Partner & Managing Director
Division Digital Analytics & Strategy