Presented By
O’Reilly + Cloudera
Make Data Work
29 April–2 May 2019
London, UK

Operationalizing risk management for machine learning

Daniel First (QuantumBlack)
Location: Capital Suite 13
Average rating: ****.
(4.00, 12 ratings)

Daniel First details QuantumBlack’s innovative methodology for identifying and mitigating risk in the development and deployment of machine learning at scale. This approach is codified in a central risk management web app, which highlights the risks associated with each step of the research and development process and provides advice to practitioners on how to overcome them. The risk management system builds on countless retrospectives, lessons learned, and “war stories” that documented the challenges analytics teams faced in the field. Join in to learn how to ensure that your machine learning models meet high-quality standards across three dimensions: performance, explainability, and fairness.

Photo of Daniel First

Daniel First


Daniel First is a data scientist at QuantumBlack. He’s worked with doctors and healthcare companies to design innovative, data-driven solutions to improve outcomes for patients by forecasting and preventing medical risks, and he’s also developed an approach to operationalize risk management for teams building machine learning models. He publishes on the social and political impact of artificial intelligence and speaks about the importance of making machine learning algorithms’ decisions interpretable to humans, most recently at the University of Oxford Mathematical Institute. He holds a BA in cognitive science and neuroscience from Yale University, an MPhil in philosophy from the University of Cambridge, where he specialized in the history of ethical thought, and an MS in data science from Columbia University.