Jan van der Vegt
The required steps to successful industrialization of machine learning models
Bringing machine learning models into production is a very different field than developing the models. Despite this, often the people building the models are responsible for industrialization. Alternatively, software engineers do not understand the intricacies of the mathematical models to connect models to the data and other systems. Collaboration and proper tooling is key to a successful machine learning project, regardless of the use case. In this talk I will go over the different steps required to industrialize models fast, flexible, secure and transparent. We will also take a look at how the Cubonacci platform helps companies to take these steps by taking over a number of these responsibilities and help with integration where customization is required.
Focus on solving complex problems using a combination of data science, data engineering, mathematical optimization, strategy consultancy and change management. Pragmatic but thorough. Competitive spirit and strong drive to keep learning. Experienced in creating a strategic vision for teams and organizations, leading a team of data scientists, coaching others and contributing to share knowledge within an organisation.
We started Cubonacci because of a clear lack of mature tooling in the data science space. Designing and building a robust technical solution like Cubonacci that captures a large enough part of the use cases has been an extremely valuable experience, both from a technical and an organizational perspective.