![](https://crypto4nerd.com/wp-content/uploads/2023/11/1C_06ssUmp7ezC44ulGkmuw.png)
In this Project I was responsible for implementing a new enterprise-wide machine learning and data science platform. This platform should enable analysts with and without deep knowledge of statistics to perform sophisticated data science tasks.
The platform was based on KNIME analytics technology. The first challenge was how to make this available throughout the company without the need for multiple on-premises servers. The solution was making use of the Microsoft Azure Cloud, enabling building machine learning solutions at scale. The Knime Servers were installed on Microsoft Windows 2019 servers and each Knime Server had several KNIME Executors attached which also were running on Windows 2019 servers. This setup enabled scaling in or scaling out the servers depending on their workload thus being cost efficient.
The second challenge was the security, how should we manage the Solution Access as well as the required connectivity to other enterprise systems? For accessing the servers, we defined Azure Active Directory Groups based on the level of access on the system as well as on organisational level. We used Graph API enabling the Single Sign On (SSO) functionality making it much more convenient for the Users to consume the interactive web-based frontend.
The connectivity to other Corporate Systems such as SAP, Snowflake, Azure Storage, Azure SQL, PowerBI and DataBricks, just to name a few, was enabled by defining Network Security Groups as well as Corporate Firewall Rules.