--- title: Uber Eats Rebuilds Search Pipeline to Cut End-to-End... description: Uber rebuilt major portions of its Eats search pipeline, achieving a 50% reduction in end-to-end latency. Key changes include switching the primary latency... canonical: https://daily.dev/posts/uber-eats-rebuilds-search-pipeline-to-cut-end-to-end-latency-by-50--f9nzd1cdw twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Uber Eats Rebuilds Search Pipeline to Cut End-to-End Latency by 50% | daily.dev og:description: Uber rebuilt major portions of its Eats search pipeline, achieving a 50% reduction in end-to-end latency. Key changes include switching the primary latency... og:url: https://daily.dev/posts/uber-eats-rebuilds-search-pipeline-to-cut-end-to-end-latency-by-50--f9nzd1cdw og:image: https://api.daily.dev/og/posts/F9nZD1CdW.png og:image:alt: Uber Eats Rebuilds Search Pipeline to Cut End-to-End Latency by 50% og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Uber Eats Rebuilds Search Pipeline to Cut End-to-End Latency by 50% **[InfoQ](https://daily.dev/sources/infoq)** · 3 min read · 5 upvotes · 0 comments ## Summary Uber rebuilt major portions of its Eats search pipeline, achieving a 50% reduction in end-to-end latency. Key changes include switching the primary latency metric to Above-the-Fold completion, trimming retrieval work with product-level embeddings, separating ranking hydration from presentation data, redesigning the advertising data path with column-oriented storage, and various infrastructure optimizations like Go garbage collection reductions. An agentic coding workflow helped identify and validate further optimizations. Engineers commenting on the work emphasized incremental, full-stack optimization over a single architectural rewrite. Uber is now exploring microbatching, product-based retrieval, Zero Pass Ranking, and HTTP multipart streaming, with early product-based search tests showing over 50% p99 latency reduction. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How did Uber Eats reduce search latency by 50%? Uber Eats cut end-to-end search latency in half by rebuilding retrieval, hydration, ranking, advertising, and infrastructure layers. Key wins included removing low-value retrieval strategies (120ms saved), product-level embeddings cutting data lookups over 100x (50ms saved), separating ranking hydration from presentation data (100ms+ saved), and redesigning advertising with column-oriented bid data and in-memory access (130ms saved). _daily.dev surfaces engineering deep dives like this for teams chasing similar latency wins._ ### What is Above-the-Fold completion as a latency metric? Above-the-Fold completion measures the time until the first screen of results is rendered with images, rather than measuring raw backend API response time. Uber Eats adopted this metric alongside pagination with server-side caching and asynchronous rendering of result items, improving perceived latency by more than 200 milliseconds compared to its prior approach. _Developers rethinking latency metrics can track architecture shifts like this on daily.dev._ ### What is Uber planning next after its search latency overhaul? Uber is exploring end-to-end microbatching, product-based retrieval, Zero Pass Ranking, and HTTP multipart streaming to let processing stages overlap instead of waiting for prior stages to finish. Early tests of product-based search retrieval have already shown more than a 50% reduction in p99 latency, suggesting further gains beyond the initial 50% end-to-end improvement. _Teams planning their own search infrastructure roadmap can follow developments like this on daily.dev._ ## Similar posts on daily.dev - [Uber Adopts Amazon OpenSearch for Semantic Search to Better Capture User Intent](https://daily.dev/posts/uber-adopts-amazon-opensearch-for-semantic-search-to-better-capture-user-intent-2caemtik8) · InfoQ · 1 upvotes · 0 comments - [Uber Improves Restaurant Recommendations Using Real-Time Signals and Listwise Ranking](https://daily.dev/posts/uber-improves-restaurant-recommendations-using-real-time-signals-and-listwise-ranking-h5fa8iuso) · InfoQ · 0 upvotes · 0 comments - [Designing a Scalable Food Delivery System Like Zomato, Swiggy and Uber Eats](https://daily.dev/posts/designing-a-scalable-food-delivery-system-like-zomato-swiggy-and-uber-eats-zozlrqfzj) · Medium · 0 upvotes · 0 comments - [How Vinted Serves Personalised Search Autocomplete](https://daily.dev/posts/how-vinted-serves-personalised-search-autocomplete-rm1zjwc0u) · Vinted · 0 upvotes · 0 comments --- Tags: [#performance](https://daily.dev/tags/performance), [#golang](https://daily.dev/tags/golang), [#distributed-systems](https://daily.dev/tags/distributed-systems), [#uber](https://daily.dev/tags/uber) [View this post on daily.dev](https://daily.dev/posts/uber-eats-rebuilds-search-pipeline-to-cut-end-to-end-latency-by-50--f9nzd1cdw) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Uber Eats Rebuilds Search Pipeline to Cut End-to-End Latency by 50%","url":"https://daily.dev/posts/uber-eats-rebuilds-search-pipeline-to-cut-end-to-end-latency-by-50--f9nzd1cdw","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/uber-eats-rebuilds-search-pipeline-to-cut-end-to-end-latency-by-50--f9nzd1cdw"},"datePublished":"2026-10-02T14:30:55.220Z","dateModified":"2026-10-02T14:31:16.342Z","description":"Uber rebuilt major portions of its Eats search pipeline, achieving a 50% reduction in end-to-end latency. Key changes include switching the primary latency...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/9266d13e9ffb08656b52d38410a7d103?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/9266d13e9ffb08656b52d38410a7d103?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"InfoQ","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"InfoQ","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/afc3bced3e1e4b188dd9127017a60e0c","url":"https://daily.dev/sources/infoq"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/uber-eats-rebuilds-search-pipeline-to-cut-end-to-end-latency-by-50--f9nzd1cdw","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":5},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"performance,golang,distributed-systems,uber","timeRequired":"PT3M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"InfoQ","item":"https://daily.dev/sources/infoq"},{"@type":"ListItem","position":3,"name":"Uber Eats Rebuilds Search Pipeline to Cut End-to-End Latency by 50%"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/uber-eats-rebuilds-search-pipeline-to-cut-end-to-end-latency-by-50--f9nzd1cdw#faq","mainEntity":[{"@type":"Question","name":"How did Uber Eats reduce search latency by 50%?","acceptedAnswer":{"@type":"Answer","text":"Uber Eats cut end-to-end search latency in half by rebuilding retrieval, hydration, ranking, advertising, and infrastructure layers. Key wins included removing low-value retrieval strategies (120ms saved), product-level embeddings cutting data lookups over 100x (50ms saved), separating ranking hydration from presentation data (100ms+ saved), and redesigning advertising with column-oriented bid data and in-memory access (130ms saved). daily.dev surfaces engineering deep dives like this for teams chasing similar latency wins."}},{"@type":"Question","name":"What is Above-the-Fold completion as a latency metric?","acceptedAnswer":{"@type":"Answer","text":"Above-the-Fold completion measures the time until the first screen of results is rendered with images, rather than measuring raw backend API response time. Uber Eats adopted this metric alongside pagination with server-side caching and asynchronous rendering of result items, improving perceived latency by more than 200 milliseconds compared to its prior approach. Developers rethinking latency metrics can track architecture shifts like this on daily.dev."}},{"@type":"Question","name":"What is Uber planning next after its search latency overhaul?","acceptedAnswer":{"@type":"Answer","text":"Uber is exploring end-to-end microbatching, product-based retrieval, Zero Pass Ranking, and HTTP multipart streaming to let processing stages overlap instead of waiting for prior stages to finish. Early tests of product-based search retrieval have already shown more than a 50% reduction in p99 latency, suggesting further gains beyond the initial 50% end-to-end improvement. Teams planning their own search infrastructure roadmap can follow developments like this on daily.dev."}}]} ```