I was a technical proof reader for "Deep-learning with Tensorflow 2 and Keras".
I worked with a financial services provider on cloud deployment of deep models for natural language processing (the proto-LLMs of 2018 such as BERT) as well as knowledge graphing. I used Tensorflow and Gremlin graph database, as well as several elements of the Amazon Web Services (AWS) stack, including Lambda serverless compute, EC2, S3, Cloud Formation, Code Pipeline, etc.
This high-level focus on practical applications of neural networks has complemented my understanding of their low-level properties.