At the second annual Women in Data Science conference, fellow and chief data officer at Western Digital Janet George spoke about industrialized data science and how the company is taking storage technology to the next level.
The conventional two-dimensional NAND flash storage architecture that allowed for very fast data access and gave birth to a proliferation of devices has reached its limits, George explained. “So we have to create a three-dimensional structure,” she said. “This three-dimensional NAND is going to get us to the skyscraper of memory technology.”
George also outlined how Western Digital uses machine learning and pattern recognition to predict how the device will perform, and to help navigate the complex journey from inventing this new type of memory technology to actually being able to mass-produce it in a factory. In part, this highlights how data science is different in manufacturing than it is in most industries, George said.
In online search or advertising, for instance, you create products and then collect data to try to understand and improve those products. “In this domain, we actually create data, and then we invent products, completely new technology nodes, to then go out and bring revenue to the company.”