The Data Standard
Building operational machine learning applications comes with many data management challenges. Feature stores are some of the most effective tools for solving them.
In this episode of The Data Standard, host Catherine Tao sits down with Willem Pienaar, a Tech Lead at Tecton, to talk about feature stores.
Feature stores are operational data systems that bridge the gap between data science and production machinery. They enable data scientists to ship offline data into production and ensure their prediction models work with accurate data.
Data is the hardest part of machine learning, as it keeps bringing chaos into an environment full of complexities that you need to simplify. Feature stores are there to help eliminate chaos and ensure you spend more time on positive rather than negative engineering.
When it comes to the future of feature stores, Willem Pienaar anticipates a greater focus on data quality. He also believes that ML data monitoring and data validation software will become more prevalent in ML and data science’s operational aspects.
Tune in with Catherine Tao to listen to this episode, as Willem Pienaar goes into detail about feature stores, data engineering and product lifecycles, challenges with ML data, and more.
Meet The Host
Data scientist at The Data Standard
Catherine Tao is a tech enthusiast looking for new methods for building connections with businesses around the world. Her extensive knowledge of data science allowed her to develop new solutions and implement them into existing ecosystems. She is currently working as a Data scientist and Exclusive Podcast Producer at The Data Standard.
Meet The Guest
Tech Lead at Tecton
Willem Pienaar is a data scientist and software engineer who frequently speaks at renowned tech conferences. He is the founder of Feast, an open-source feature store and is currently working as a Tech Lead at Tecton.