Add advanced topics section

[rmarquis]

We have left out quite a few topics aside due to lengths of the course to focus on the core ML experiences.
Adding an advanced section (or adding a few chapter in the "conclusion" part) might be ideal to explore a few advanced topics, such as:

• distributed annotation with LabelStudio and sync before retraining
• self-hosted runner with Dockerization of the trained model using tools like KubeVirt and Kaniko
• explainability (see the code branch with an attempt with GRAD cam)
• testing
• observability (usage and inference performance with BentoML)
• VPNs
• Cache dependencies of runner
• Infrastructure as Code IaC with Terraform
• Feature store (Feast)
• Streaming vs Batch prediction
• Stream images from buckets (instead of complete pull)
• standardized Python version