Kinesis Data Analytics with MSK and Lambda

Hi, We are exploring with KDA for our streaming data ingestion pipeline for ML purpose. From https://aws.amazon.com/blogs/machine-learning/using-streaming-ingestion-with-amazon-sagemaker-feature-store-to-make-ml-backed-decisions-in-near-real-time/ it seems possible to connect KDA with MSK and use Lambda as destination to ingest data to Sagemaker Feature Store (or any custom location). However checking with KDA documentation it seems connecting with MSK is only possible with KDA studio or as a Flink app which does not provide a Lambda destination directly - instead would require another Kinesis stream in between, whereas KDA SQL application does support connecting with Lambda but not support reading from MSK. Our use case is very similar to the blog post above and we prefer to use Python as much as possible. What can be a viable path here?

Thanks, Rob