September 2025 Summaries

E-commerce is evolving into a domain where personalized, dynamic, and reliable experiences are paramount, and Vespa is positioned as a key player in facilitating this transformation. By integrating advanced search and recommendation capabilities with machine learning, Vespa offers a unified platform that prioritizes real-time personalization, multi-objective optimization, and efficient data handling. This enables businesses to deliver tailored user experiences while balancing cost, innovation speed, and scalability. Vespa's ability to support seamless collaboration across various functional teams and its capacity for real-time updates and granular control make it a preferred choice for global leaders in retail, media, and finance. Industry analysts and large corporations recognize Vespa's capabilities in enhancing search and personalization, contributing to its adoption at scale for billions of daily queries. As e-commerce shifts towards creating personalized journeys for users, Vespa stands out by offering a comprehensive solution that aligns with both user expectations and business goals.

As AI applications advance, the limitations of vector databases, such as their inability to fully capture complex relationships across various data modalities, are becoming evident, which is where tensors provide a more robust solution. While vector databases excel at fast retrieval through approximate nearest neighbor search, they often lack full-text search capabilities, integration with structured data, and support for custom ranking, leading to inefficiencies in applications requiring personalization and real-time updates. Tensors, as multi-dimensional numerical representations, preserve context and relationships, enabling more precise and explainable retrieval tasks, such as hybrid logic and multimodal understanding. Vespa's tensor system offers a scalable, expressive framework that supports both dense and sparse data dimensions, with strong typing and a minimal set of operations, allowing for seamless integration of symbolic and semantic search. This approach not only enhances the ability to reason with data but also provides a scalable, high-performance platform suitable for real-time AI applications.

The Vespa Newsletter for September 2025 highlights significant enhancements and new features in Vespa's AI and data management capabilities. Key updates include the introduction of Approximate Nearest Neighbor (ANN) tuning parameters in their HNSW implementation, which offer greater control over recall and performance trade-offs, and automatic binary data detection in string fields to ensure data integrity in GenAI pipelines. The newsletter also covers improvements in filtering and grouping, such as the ability to apply logical predicates and range filters, and introduces the geoBoundingBox operator for geographical filtering. Enhancements to Pyvespa facilitate easier application development and deployment directly from Python, and the newsletter notes that Vespa Cloud will now feature automatic instance migration to new hardware generations for improved performance and cost efficiency. Additionally, the update discusses the benefits of layered ranking and multiple inheritance of summary features for more flexible ranking strategies, and encourages users to engage with the Vespa community for further exploration and feedback.

Jan Böker's blog post provides insights into optimizing Vespa's Approximate Nearest Neighbor (ANN) search parameters, particularly with the introduction of ACORN-1 in the Hierarchical Navigable Small World (HNSW) algorithm. The post discusses different strategies Vespa employs for ANN searches with filters, including pre-filtering HNSW, post-filtering, and the newly introduced ACORN-1 strategy. It details how Vespa automatically chooses a search strategy based on filter hit ratios and how adjusting parameters like filter-first-threshold and approximate-threshold can influence performance, response time, and recall. The blog emphasizes the importance of tweaking these parameters to balance response time and recall, noting that changes in parameters can lead to different outcomes based on the dataset's characteristics. The discussion also touches on using parameters like exploration-slack and exploreAdditionalHits to fine-tune recall and response time further, highlighting the trade-offs between increased recall and response time.

Jan Böker, a software engineer, discusses recent enhancements to Vespa's HNSW (Hierarchical Navigable Small World) algorithm used for approximate nearest-neighbor (ANN) search in applications like search and recommender systems. The blog post highlights the addition of ACORN-1 and adaptive beam search to improve filtered ANN search, which allows for query-time constraints. ACORN-1 aims to optimize the search by exploring 2-hop neighbors first, reducing unnecessary computations, while adaptive beam search provides a distance-based termination condition to improve recall with fewer distance computations. These enhancements are tested and compared to previous implementations, showing significant improvements in search performance, with ACORN-1 particularly effective in reducing response time without compromising recall. The post also explores the balance between response time and recall when implementing multi-hop neighbor exploration and discusses the implications of these innovations for future search strategies in Vespa.

Onyx, an open-source AI platform designed for enterprise knowledge management, successfully utilized Vespa Cloud's automated configuration changes to achieve a 24.5% cost reduction without sacrificing performance or requiring extensive manual intervention. As Onyx's customer base expanded, managing their Vespa deployment became increasingly complex, necessitating a balance between cost and performance. Vespa Cloud offered insights and suggestions for resource optimization, allowing Onyx to reconfigure their system efficiently. This included reducing CPU overprovisioning and optimizing memory usage, which were facilitated by Vespa Cloud's automated deployment and migration capabilities, ensuring a seamless transition with no service interruptions. The platform's tools and expert support enabled Onyx to continuously optimize their infrastructure as they scale, setting the stage for further improvements in efficiency and cost savings.