Home

Machine learning applications are ubiquitous today, but the development of those applications is often ad-hoc and lacks reproducibility or scalability. With machine learning at the core of Digits, we have focused on automating our processes around model training and deployments. In this talk, Jo & Hannes will share their key learnings and insights into the machine learning tools used at Digits. They’ll also discuss the benefits of focusing on ML Engineering besides the actual model development work. You’ll gain an understanding on how to take a machine learning experiment, build machine learning pipelines and deploy it into a production environment.

The practice of data science requires data — data that comes in all forms, formats, platforms and structures. Because of this complexity we commonly hear that data scientists spend as much as 80% of their time ‘wrangling’ data and only 20% producing actionable outcomes. We invite you to attend this session to see how the Pareto rule can be flipped by a demonstration of data virtualization tools within data science notebooks, using Python and R.

In a technology-heavy industry, how do you actually present the value of technology – in this case data science – to those that are surrounded by it, but don’t understand it? In this talk, I’ll present some useful tips for communicating the what and why of data science to a non-technical audience using examples from when I’ve had to do it! I’ll then present the AI product that we built for radiologists at Radiology Partners and discuss some challenges we are facing in operationalizing such a solution in clinical practice.

Many companies struggle with data science project development and deployment. In these cases, the majority of data scientists’ time is often spent on data exploration and data preparation. Data processing is time consuming and may even require months to re-code the model in the production system for deployment. In this talk, you will learn how Alliance Data has accelerated data science. By leveraging big data engineering, Alliance Data is able to streamline data acquisition, build out repeatable and scalable data ingestion and curation pipeline, automate feature engineering and use a variety of tools to speed up model development and deployment.

In this workshop, we will introduce you to TensorFlow 2.0 and to model-building styles for beginners and experts, including sequential, functional, and subclassing APIs. You’ll see complete, end-to-end code examples in each style, from “Hello, world” all the way through advanced examples. We will also demonstrate how to deploy your trained models using TensorFlow Serving to support real-time and batch inference.

Advances in data and analytics have the potential to transform the organizations that deliver and pay for healthcare. Despite this promise, barriers endemic to the industry have made healthcare organizations slow to adopt and realize the benefits of these innovations. In this talk, Shannon Hollars will discuss these challenges facing healthcare, the tremendous possibilities of data and analytics, and why analysts and data scientists can find this an attractive and deeply rewarding field in which to contribute and sharpen their skills.

Although we’re most familiar with Uber’s surge pricing, dynamic pricing actually has a long history across retail, transportation, hospitality, professional sports, and other industries. In this presentation, Rob will lead a discussion of dynamic pricing, focusing on how casinos are embracing model this to optimize real-time pricing for slot machines, and how the concept might be applied to other industries.

This talk will focus on the intersection of data science and genomics research. Have you ever wondered about the research that allows Ancestry or 23andme to tell your traits just looking at your genome? Or how we can use genetic information to advance treatments for complex conditions like diabetes, heart abnormalities, Parkinson disease, and Crohn disease? Well, that’s what Ezgi is aiming to tell us about. She will also talk about an R package (gwasurvivr) she co-developed in her lab, and why having data science skills is so crucial for biomedical research in the post-genomic era.