January 2023 Summaries

Feature monitoring for real-time machine learning is crucial to prevent "garbage in, garbage out" situations and ensure model accuracy. It involves monitoring feature values after transformation, as well as aggregations of features such as metrics or statistics. Common obstacles include volatile dependencies on analytics teams, computing and validating feature metrics, limitations of current tools, and understanding and detecting data drift. These challenges can lead to inaccurate predictions and suboptimal performance if not addressed. Feature monitoring requires regular assessment of data quality, handling temporal joins and align timestamps, performing metric compute at scale, and implementing validation rules that take action when necessary. Addressing data drift is a complex process that often requires advanced algorithms and visualizations.

A feature platform is a crucial component of the modern machine learning (ML) stack, enabling companies to build and deploy ML models quickly and reliably. By standardizing ML data workflows from design to production, a good feature platform helps break down silos, reduces tech debt, and increases efficiency for ML teams. This leads to improved reliability, reduced infrastructure costs, and the ability to unlock new revenue sources with real-time data. A well-designed feature platform should enable ML teams to manage and improve their features efficiently, while minimizing associated costs, and be compatible with a wide range of tools and technologies. By adopting a feature platform, companies can reap multiple benefits, including increased efficiency, improved reliability, and the potential for new revenue streams through real-time data utilization.

As Head of Growth at Tecton, Claire Besset's team focuses on making the platform more widely known to potential users in the machine learning community. They optimize their website, host events, and create resources such as the apply() virtual conference to share with the community. Before joining Tecton, Claire held a similar role at another startup that sold a machine learning solution for B2C marketing teams. She was attracted to Tecton because of its potential in the growing feature stores and MLOps category. Claire is a fan of the Notion tool, which she uses to manage her growth marketing roadmap. In her free time, she enjoys practicing "stick and staff fighting," a French martial art that emphasizes elegance over efficiency. She recently read "Mathematica" by David Bessis, a book about how mathematicians think about mathematics and the human mind. During the pandemic lockdown, Claire lived in Brooklyn with her husband, who moved into her apartment as a temporary solution until they could move into their new home.