This project is intended for the IAM (Identity and Access Management) space. We currently use IMA as our IAM tool. In an IAM tool you can for example request access to an application, another person can then accept or reject your request, and the tool will then implement that access, for example in active directory.
The data set in IMA is large; there are more than 30,000 users and more than 1,000,000 roles in Proximus.
Currently, we are migrating IMA to a new platform called BFARM. This is a big project that will take us a long time to complete. Because of the long timespan both solutions IMA and BFARM need to run in parallel. Because people can manage rights in 2 applications, there is a difference between the rights granted by the two applications.
You would participate in this project by computationally comparing and analyzing the differences in user rights across three tools: IMA (legacy), BFARM (new) and Active Directory (actual implementation) within the Proximus environment.
Implementation
The tool should get the raw data from all three tools and convert it to a standardized data structure such as a 3D tensor to allow for ease of comparison across tools.
A GUI should allow users to navigate, filter, and generate reports on discrepancies and/or anomalies.
Beyond this, we anticipate that there should be risk-based filtering to prioritize high-risk roles or users for review. This, however, could be outside of the scope of the internship if it proves to be challenging within the timeframe.
The solution should be built in Azure for efficient deployment and integration.
Requirements
Strong knowledge of Python programming.
Experience or willingness to learn Azure platform and Azure ML pipelines.
Basic understanding of linear algebra for matrix operations.
Optional: Front-end development skills for GUI implementation.
Interest in computationally complex data problems and IAM systems.