Twitter has rapidly evolved in how we do Machine Learning over the last few years. We rapidly evolved from making incremental improvements to our existing Logistic Regression models to a wave of new Deep Learning architectures. At the same time, we were faced with the challenge of uneven distribution of ML expertise across the company. This is the story of the Cortex team that enabled this machine learning transformation across the entire company by making a set of careful strategic investments.

Agile practices has become the default choice for software delivery. However, developing different types of software requires different approach and techniques. For example - a website development is different from developing a data lake, a mobile application development is different from an IOT or Robotics Automation project. In this session, we would focus on how knowing the domain of development can ease out your team delivery challenges and how we can embed the domain of development into agile practices. We will take example of a data lake project to demonstrate how a data lake development project can be delivered in a agile fashion.