{"id":151,"date":"2026-11-17T05:10:05","date_gmt":"2026-11-17T05:10:05","guid":{"rendered":"https:\/\/blockoala.com\/index.php\/2026\/11\/17\/uber-eats-search-overhaul-cut-end-to-end-latency-by-50-infoq-reports\/"},"modified":"2026-11-17T05:10:05","modified_gmt":"2026-11-17T05:10:05","slug":"uber-eats-search-overhaul-cut-end-to-end-latency-by-50-infoq-reports","status":"publish","type":"post","link":"https:\/\/blockoala.com\/index.php\/2026\/11\/17\/uber-eats-search-overhaul-cut-end-to-end-latency-by-50-infoq-reports\/","title":{"rendered":"Uber Eats search overhaul cut end-to-end latency by 50%, InfoQ reports"},"content":{"rendered":"<p><em>Originally reported by InfoQ on November 16, 2026.<\/em><\/p>\n<p>Uber rebuilt major parts of the Uber Eats search pipeline, cutting end-to-end search latency by 50%, according to <a href=\"https:\/\/www.infoq.cn\/article\/ah2izwCUPxX2y1herN7b\" rel=\"noopener nofollow\" target=\"_blank\">InfoQ\u2019s report<\/a>. The changes span retrieval, feature hydration, ranking, ads, rendering and infrastructure rather than a single architectural overhaul.<\/p>\n<p>Uber changed what it measured. Instead of focusing only on backend API response time, it tracked <strong>above-the-fold completion<\/strong>: the time from a request until the first screen of results, including images, is fully rendered. Server-side caching for pagination shortened the initial response, while asynchronous rendering allowed result items to be processed concurrently. Uber reported that these changes reduced above-the-fold latency by more than 200 milliseconds.<\/p>\n<p>The team also found that tens of thousands of candidates were receiving feature data before ranking, even though many would later be discarded. Removing low-value retrieval strategies saved about 120 milliseconds. Product-level embeddings reduced the number of data queries by more than 100-fold and saved another 50 milliseconds. Separating ranking features from display data cut latency by more than 100 milliseconds, while removing unnecessary dependencies and request hedging saved 35 and 40 milliseconds, respectively.<\/p>\n<p>In the ads path, column-oriented bidding data, in-memory access and less serialization reduced latency by about 130 milliseconds. Other infrastructure changes included parallel encoding, smaller embedding vectors, connection-management improvements and Go data-structure changes to reduce garbage-collection overhead. Uber also introduced an agentic-coding workflow to identify, benchmark and validate further optimizations.<\/p>\n<p>Uber is exploring end-to-end microbatching, product-based retrieval, Zero Pass Ranking and HTTP multipart streaming. According to the report, early tests of product-based retrieval reduced <strong>p99 latency by more than 50%<\/strong>. That early-test result is distinct from the reported 50% reduction in end-to-end search latency; the proposed pipeline changes remain under exploration.<\/p>\n<p>Source: <a href=\"https:\/\/www.infoq.cn\/article\/ah2izwCUPxX2y1herN7b\" rel=\"noopener nofollow\" target=\"_blank\">infoq.cn<\/a> (originally published 2026-11-16 17:24 UTC+8)<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Originally reported by InfoQ on November 16, 2026. Uber rebuilt major parts of the Uber Eats search pipeline, [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":150,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-151","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-uncategorized"],"_links":{"self":[{"href":"https:\/\/blockoala.com\/index.php\/wp-json\/wp\/v2\/posts\/151","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/blockoala.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/blockoala.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/blockoala.com\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/blockoala.com\/index.php\/wp-json\/wp\/v2\/comments?post=151"}],"version-history":[{"count":0,"href":"https:\/\/blockoala.com\/index.php\/wp-json\/wp\/v2\/posts\/151\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/blockoala.com\/index.php\/wp-json\/wp\/v2\/media\/150"}],"wp:attachment":[{"href":"https:\/\/blockoala.com\/index.php\/wp-json\/wp\/v2\/media?parent=151"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/blockoala.com\/index.php\/wp-json\/wp\/v2\/categories?post=151"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/blockoala.com\/index.php\/wp-json\/wp\/v2\/tags?post=151"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}