The ML Platforms track focuses on the practice of moving machine learning systems from development to deployment, and the next-generation tooling that makes this an organic process. Removing barriers to deployment allows data scientists to gain quicker feedback and deeper insights into how models are performing in the wild, and enables nimbler experimentation at the product level.
To do this, new machine learning workflows are taking advantage of serverless and container orchestration technologies, and also specializing frameworks towards certain classes of data science problems.
This track is for the hands-on practitioners who fluidly cross the boundary between research and deployment.
Prasanna Padmanabhan leads the ML infrastructure for Recommendations team at Netflix. His primary focus is to make it easy for Researchers to innovate faster and reduce the time it takes from ideation to production. His main challenge is to scale ML in heterogeneous language environments across several domains and at all stages of a model lifecycle, including ad-hoc exploration, preparing training data, model development, and robust production deployment. In the past, he has built high scale distributed data systems that leverages both batch and stream processing.