PaperCall is shutting down on August 31, 2026. Existing events and submissions will remain available until then. New events can no longer be created.

Serverless Deep Learning with TensorFlow and Serverless Framework

By Pratik Parmar

Elevator Pitch

Deep and machine learning is becoming essential for a lot of businesses. But one of the main issues companies face with deep/machine learning is finding the right way to deploy these models. Serverless provides a cheap, scalable and reliable architecture for deep learning models.

Description

During this session we’ll cover the following: - Using serverless for deep learning - standard ways of deploying deep learning applications, and how a serverless approach can be beneficial.

  • “Hello world” code - a basic Lambda function with only 4 lines of code. There is no API here, we’ll start with the simplest possible example.

  • Code decomposition - looking through the configuration file, and the python code for handling the model, to understand how the whole example works.

  • API example - get a working API for image recognition on top of our example.