Building a Production-Grade Vector Database in Rust: What We Shipped

<!-- SC_OFF --><div class="md"><p>Hi Reddit, we are a team of database researchers (including a PhD from MIT DB Group) and we just open-sourced an embedded vector database for

<p><a class="article-body-image-wrapper"

<p>We paid for a vector database for a year. It was faster than Postgres at the one thing it does. We deleted it anyway.<br /> </p> <div class="highlight js-code-highlight"> <pre

<h1> Moving Beyond Simple Vector Search: Why Hybrid Search is Essential for RAG </h1> <p>As LLMs continue to dominate the landscape, Retrieval-Augmented Generation (RAG) has

<!-- SC_OFF --><div class="md"><p>I ported NGT (Yahoo Japan's ANN library) to Rust, then implemented TurboQuant compression and attempted GPU acceleration via Metal. Here's what

<p>I have helped a number of teams pick a vector database in the last year. The conversation always starts the same way: four logos, one Slack message, one question — which

<p>The chunk was in the index. I checked three times.</p> <p>A support bot kept answering "that isn't covered in the documentation" for a question that was covered, in one

<p>Most learning-to-rank examples end when the model has been trained. A production search system still has several harder questions to answer:</p> <ul> <li>Which user events are