<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Infino Blog</title><description>Engineering notes and anonymized case studies from the Infino team.</description><link>https://infino.ai/</link><item><title>The one idea behind our speed on Parquet</title><link>https://infino.ai/blog/the-invention-behind-our-parquet-speed/</link><guid isPermaLink="true">https://infino.ai/blog/the-invention-behind-our-parquet-speed/</guid><description>One searchable copy instead of three systems, the idea behind fast search on Parquet, plus where Solr still wins.</description><pubDate>Wed, 01 Jul 2026 00:00:00 GMT</pubDate><category>engineering</category></item><item><title>How a full-text and vector index fits inside a valid Parquet file</title><link>https://infino.ai/blog/search-index-inside-a-parquet-file/</link><guid isPermaLink="true">https://infino.ai/blog/search-index-inside-a-parquet-file/</guid><description>The byte layout: BM25 and vector indexes between the last row group and a rewritten footer, still readable by DuckDB.</description><pubDate>Wed, 01 Jul 2026 00:00:00 GMT</pubDate><category>engineering</category></item><item><title>Why we added a search index to Parquet instead of designing a new format</title><link>https://infino.ai/blog/why-parquet-not-a-new-format/</link><guid isPermaLink="true">https://infino.ai/blog/why-parquet-not-a-new-format/</guid><description>Columnar formats are bad at search, so we added an index layer instead of replacing Parquet, including where that hurts.</description><pubDate>Wed, 01 Jul 2026 00:00:00 GMT</pubDate><category>engineering</category></item><item><title>How we built the engine: files, manifests, and stateless compute</title><link>https://infino.ai/blog/how-we-built-the-engine/</link><guid isPermaLink="true">https://infino.ai/blog/how-we-built-the-engine/</guid><description>Files, immutable manifests, and stateless compute: how the engine is assembled over object storage.</description><pubDate>Wed, 01 Jul 2026 00:00:00 GMT</pubDate><category>engineering</category></item><item><title>How full-text, BM25, and vector search actually work</title><link>https://infino.ai/blog/how-full-text-and-vector-search-work/</link><guid isPermaLink="true">https://infino.ai/blog/how-full-text-and-vector-search-work/</guid><description>Inverted indexes, BM25, IVF, and rank fusion, explained through how we actually built them.</description><pubDate>Wed, 01 Jul 2026 00:00:00 GMT</pubDate><category>engineering</category></item><item><title>Making S3 the source of truth: the tricks that are not in the docs</title><link>https://infino.ai/blog/s3-source-of-truth/</link><guid isPermaLink="true">https://infino.ai/blog/s3-source-of-truth/</guid><description>Two-GET opens, a one-CAS commit, and a cache that drops RAM but keeps the file on disk.</description><pubDate>Wed, 01 Jul 2026 00:00:00 GMT</pubDate><category>engineering</category></item><item><title>Lessons from building a search engine in Rust</title><link>https://infino.ai/blog/lessons-building-search-in-rust/</link><guid isPermaLink="true">https://infino.ai/blog/lessons-building-search-in-rust/</guid><description>Why tokio drives I/O and rayon drives CPU, and the 3x to 10x cold-search regression that taught us the rule.</description><pubDate>Wed, 01 Jul 2026 00:00:00 GMT</pubDate><category>engineering</category></item><item><title>We swapped Solr out of a real search app. Here are the numbers.</title><link>https://infino.ai/blog/swapping-solr-for-infino/</link><guid isPermaLink="true">https://infino.ai/blog/swapping-solr-for-infino/</guid><description>BM25 + vectors + facets from one Parquet file, including where Solr still wins.</description><pubDate>Wed, 01 Jul 2026 00:00:00 GMT</pubDate><category>engineering</category></item><item><title>License-usage analytics for a design-automation startup</title><link>https://infino.ai/blog/license-analytics-eda/</link><guid isPermaLink="true">https://infino.ai/blog/license-analytics-eda/</guid><description>A design-automation startup puts 20 years of manual EDA license analysis into queryable form.</description><pubDate>Wed, 01 Jul 2026 00:00:00 GMT</pubDate><category>case study</category></item><item><title>Product analytics at 25 billion events a day</title><link>https://infino.ai/blog/product-analytics-25b-events/</link><guid isPermaLink="true">https://infino.ai/blog/product-analytics-25b-events/</guid><description>Why a large consumer mobile app is indexing its event stream in place on S3.</description><pubDate>Wed, 01 Jul 2026 00:00:00 GMT</pubDate><category>case study</category></item><item><title>A canonical people index for legal-tech</title><link>https://infino.ai/blog/canonical-people-index/</link><guid isPermaLink="true">https://infino.ai/blog/canonical-people-index/</guid><description>How a legal-tech company models a canonical people index on hybrid search.</description><pubDate>Wed, 01 Jul 2026 00:00:00 GMT</pubDate><category>case study</category></item><item><title>A staging area for thirty years of court records</title><link>https://infino.ai/blog/court-records-staging-area/</link><guid isPermaLink="true">https://infino.ai/blog/court-records-staging-area/</guid><description>A government records modernization uses Infino as the enrichment layer between two systems of record.</description><pubDate>Wed, 01 Jul 2026 00:00:00 GMT</pubDate><category>case study</category></item></channel></rss>