{"en":"This page is not currently available in your language.","id":"Halaman ini sekarang belum tersedia dalam bahasa Anda.","ms":"Laman ini kini belum boleh didapati dalam bahasa anda.","es_ES":"Esta página no está disponible en tu idioma en este momento.","da_DK":"Denne side er i øjeblikket ikke tilgængelig på dit sprog.","fr":"Cette page n'est actuellement pas disponible dans votre langue.","de":"Diese Seite ist zurzeit nicht in Ihrer Sprache verfügbar.","en_GB":"This page is not currently available in your language.","pl":"Ta strona aktualnie nie jest dostępna w Twoim języku.","it":"Al momento questa pagina non è disponibile nella tua lingua.","nl_NL":"Deze pagina is momenteel niet beschikbaar in je taal.","nb_NO":"Denne siden er for øyeblikket ikke tilgjengelig på språket ditt.","ru":"На данный момент эта страница недоступна на вашем языке.","zh_TW":"此頁面目前沒有您語言的版本。","sv_SE":"Denna sida är för närvarande inte tillgänglig på ditt språk.","th_TH":"หน้านี้ยังไม่มีให้บริการในภาษาของคุณในขณะนี้","zh_CN":"本页面暂无您所需的语言版本。","ja":"申し訳ありませんが、このページは選択した言語ではご利用になれません。","pt_BR":"Esta página não está disponível no momento em seu idioma.","ko":"현재 이 페이지는 한국어 번역이 제공되지 않습니다.","es":"Esta página aún no está disponible en tu idioma.","uk_UA":"На даний момент ця сторінка недоступна на вашій мові.","en_AU":"This page is not currently available in your language."}
September 27, 2018

The architecture behind Nautilus, our new search engine

Filed under: Technology

Search presents a particular challenge for Dropbox with our massive scale. Our users keep hundreds of billions of pieces of content on our collaboration platform, and we need to provide each user with a personalized search experience. Our Intelligence Group’s Search Infrastructure team rebuilt our search function from the ground up through a project called Nautilus. Along with delivering best-in-class performance, scalability, and reliability, they also provided a foundation for implementing intelligent document ranking and retrieval features. This is a flexible system that allows our engineers to easily customize the document-indexing and query-processing pipelines for running experiments. And, of course, Nautilus also maintains strong safeguards for preserving the privacy of our users’ data.

“We want to help users find the most relevant documents for a given query—at this particular moment in time—in the most efficient way possible,“ writes engineering lead Diwaker Gupta in today’s post on our tech blog. “This requires being able to leverage machine intelligence at several stages in the search pipeline, from content-specific machine learning (such as image understanding systems) to learning systems that can better rank search results to suit each user’s preferences.”

Update: A second Nautilus post focuses on performance and reliability.