Problem: Data Science teams are continuously building and traning models, this quick and constant iteration has created stress on the infrastructure. There are many problems with the exisiting system: it does not scale to multiple regions easily, low observability in the system/model, poor model management, unused compute resources, and many more issues.

Your task: is to design and implement a system that address the challenges Data Science teams face. YOU ARE NOT limited to the challenges mentioned above, do your research and make your system as robust as possible in the given time. However you must satisfy the following basic requirements:

• Two availability zones
• Basic infra & model monitoring
• Leverage CI/CD to test and implemtent your code
• You should be able to load and run a test data model
• NO KUBEFLOW

Finally you will give a 10 minute demo/presentation about your solution and the problems it solves, followed by a 5 minute Q&A. Architecture diagram is a must.

This is intentionally vauge, you will be assessed on your decision making and systems design. The winning team will be imortialized in this repo.

Winning Teams:

• 20B: