Machine Learning Inference at Scale with Python and Stream Processing
There is frequently an “impedance mismatch” between developing and training a machine learning model (a data scientist’s job) and then deploying that model to perform at scale in a production environment (a data engineer’s job). How do you make a trained prediction model usable in real time, while the user is interacting with your software? What does it take to go from fast trial-and-error runs on historical data to models that perform at production scale, in real time?
In this talk we will show you how to write a low-latency, high throughput distributed stream processing pipeline (in Java), using a model developed in Python.