Panel Discussion: Managing Model Risk
11:00 - 11:30 | Friday, May 24
The panel discussion will address the topic of managing the risks associated with the development, implementation and use of models in decision-making. Predictive models have a lifecycle, and before addressing each phase separately, the panel will provide a brief overview of each stage. Some specific model risks will be discussed at each phase, such as:
- Model Performance Monitoring: model degradation for models that need to be retrained on a set cadence;
- Model Development: trade-off between model complexity and model performance, model bias towards protected class proxy variables, model explainability, usage of pre-trained models in NLP projects;
- Model Implementation: model obsolescence or getting a model to production fast enough once it’s trained
Director, Data Science, USAA
Silvia Ochoa is currently serving as the leader of the Data Science practice in the USAA Federal Services Bank. The function is responsible for providing data science and machine learning solutions that enhance bank business processes and member experiences. Previously, she served in data and analytics leadership roles in the USAA Marketing and Chief Risk Office (Model Risk Management) organizations. Silvia holds a Ph.D. in Applied Mathematics from University of Wyoming and a M.Sc. degree in Numerical Calculus and Statistics from Babes-Bolyai University, Romania.
Head of Analytics & Data Strategy, Kabbage
Technology entrepreneur and data science leader with successful track record of developing analytical systems, teams, and businesses from the ground up.
Co-Founder of Orchard, a pioneering data, analytics, and transaction platform that accelerated the growth and institutionalization of Online Lending during a time of massive scale. Acquired by Kabbage in 2018.
Frequent writer and speaker on financial technology, consumer credit, and the future of data-driven business and its impact on society.
Data Scientist, Multinational Finance Firm
Mariem’s background is in computer science and engineering. She started her career as a software developer. She later discovered a passion for data science. Mariem works on various data sets and her projects touch upon BI, text mining, automation, sentiment analysis and recommendation systems. In addition to her day-to-day role, Mariem has a strong interest in tech for social good; she at times volunteers at hackathons for non-profits. Mariem is also an advocate for getting more people into technology: she has previously created and taught Python and web programming workshops; she has also been featured in the book, “Her STEM Career: Adventures of 51 Remarkable Women”, aimed at introducing girls to various STEM fields.