Lance Hokanson, Senior Business Intelligence Manager @ F5 Networks

Traditional data warehouse & data engineering pipelines worked from the bottom up; figure out what data you need, connect to the data, transform and load the data. This model worked great for architectures where bandwidth was narrow, latency was high, and database speeds were slow. Using modern data engineering and data lake platforms you can turn the model on its head and load everything upfront, provide quicker access to broad swaths of data to analysts and scientists, then spend your precious data engineering time curating only what you need to operationalize.