Reza Rokni
Developer Advocate, Google
In his 7 years at Google, Reza has been lucky enough to work with developers from many industries from Gaming to Banking, applying Google’s Data Analytics technologies to new domains. Currently, as a developer advocate for Google Cloud Dataflow and Apache Beam, he gets to have fun talking about stream processing all day long, or should that be stream processing unbounded… 🙂
Past sessions
In this session we will be using an LSTM Encoder-Decoder Anomaly Detection model as an example, to show the building and retraining of a model which uses the tfx-bsl package to run continuous inference. We will also emphasize the importance of the hermetic seal between training and inference paths.
Speakers: Robert Crowe and Reza Rokni
