add two new exercises

notebooks/en/mlflow_ray_serve.ipynb

@@ -1055,6 +1055,8 @@
         "## Exercises\n",
         "\n",
         "- The generated API signature is **very similar** to the model signature, but there's still some mismatch. Can you identify where it is? Try fixing it. Hint: What happens when you try passing in multiple questions to the question-answerer endpoint we set up?\n",
+        "- MLflow model signatures allow for [optional inputs](https://mlflow.org/docs/latest/model/signatures.html#required-vs-optional-input-fields). Our current implementation does not account for this. How might we extend the implementation here to support optional inputs?\n",
+        "- Similarly, MLflow model signatures allow for non-input [\"inference parameters\"](https://mlflow.org/docs/latest/model/signatures.html#model-signatures-with-inference-params), which our current implementation also does not support. How might we extend our implementation here to support inference parameters?\n",
         "- We use the name `DynamicallyDefinedDeployment` every single time we generate a new deployment, regardless of what model name and version we pass in. Is this a problem? If so, what kind of issues do you foresee this approach creating? Try tweaking `deployment_from_model_name()` to handle those issues."
       ]
     }