QSU Research Methods Seminar: David Berg, PhD

QSU Research Methods Seminar
Tuesday, May 12, at 12:00 pm PST
Join the seminar at Tapao Hall (3180 Porter Dr. combined Room B107 & B136)
*Lunch will be provided*

David Berg, PhD
Senior Software Engineer at Netflix
Ph.D in Computational Neuroscience
Reliability and Reproducibility in machine learning systems

Modern scientific research, whether for companies like Netflix, academia or statistical medicine, rely on a variety of complex infrastructure ranging from access to data, compute, lineage, versioning, workflow orchestration and scheduling. Although important, these concerns are often an impediment to applied scientists. This talk will focus on how the machine learning platform at Netflix uses infrastructure tools like Metaflow, our open source infrastructure, to support reliable and reproducible experiments and ML systems at scale without burdening scientists. With our infrastructure a small team can support hundreds of use cases across many domains such as product experiments, causal inference, machine vision, predictive modeling and generative AI. As motivation for the infrastructure choices we’ll look at several case studies across several important business domains.

Zoom Link:
Meeting ID: 977 5911 6610
Password: 746758
We welcome any suggestions or questions about the seminar series – please contact Ni Deng nideng@stanford.edu
We look forward to seeing you!
http://transportation.stanford.edu/marguerite/
http://transportation.stanford.edu/marguerite/MargueriteSched.shtml