April 2024 Summaries

At the annual Stripe Sessions conference held at the Moscone Center in San Francisco, Stripe unveiled significant updates and advancements across its core product areas, emphasizing its evolving interoperability with other payment providers. Notable announcements included enhancements to the Global Payments suite, such as an expanded Optimized Checkout Suite offering over 100 payment methods and advanced features like A/B testing and Adaptive Pricing. Stripe's Terminal product now supports enterprise features and has expanded its integration with various POS and PMS providers. The Advanced Payments Engine has been upgraded with machine learning-based optimizations, while new fraud prevention tools like Radar Assistant were introduced. Stripe announced the launch of Organizations for managing multiple accounts and significant updates to Stripe Connect, which now offers modular components and pricing tools. The Embedded Finance section saw new components for Treasury, Issuing, and Capital, alongside new fraud tools. The Revenue and Finance Automation platform introduced features like the Meters API for usage-based billing and advanced discounting logic. Stripe also highlighted its efforts to integrate with third-party providers, launching new APIs and expanding its app marketplace to better fit into users' existing technology stacks.

Machine learning (ML) is deeply integrated into Stripe's operations, enhancing various processes and contributing significantly to the global internet economy. One of the main challenges at Stripe is feature engineering, the iterative process of defining inputs for ML models, which is complicated by the vast amount of raw data. To address this, Stripe partnered with Airbnb to adapt its Chronon platform, creating Shepherd, a next-generation ML feature engineering platform that meets the company's strict requirements for latency and feature freshness. This platform has been instrumental in developing a new fraud detection model with over 200 features, which has significantly reduced fraud losses. As Stripe continues to refine Shepherd, it is contributing enhancements to Chronon, supporting its open-source community, and integrating the platform with Stripe's data systems to handle large-scale offline and streaming data requirements. The collaboration with Airbnb and the development of Shepherd exemplify Stripe's commitment to advancing ML infrastructure and open-source contributions.