<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Furkan Turkoglu — Blog</title><description>Blog posts on backend systems, PostgreSQL/pgvector hybrid search, and applied AI.</description><link>https://portfolio.furkanturkoglu.workers.dev/en/</link><language>en-us</language><item><title>Why vector similarity alone falls short: PostgreSQL FTS, pgvector, and RRF</title><link>https://portfolio.furkanturkoglu.workers.dev/en/blog/hybrid-retrieval-postgresql-pgvector-rrf/</link><guid isPermaLink="true">https://portfolio.furkanturkoglu.workers.dev/en/blog/hybrid-retrieval-postgresql-pgvector-rrf/</guid><description>Why dense embeddings miss exact technical tokens, and how to fuse PostgreSQL full-text search with pgvector cosine similarity using RRF in SQL.</description><pubDate>Fri, 18 Sep 2026 00:00:00 GMT</pubDate><category>postgresql</category><category>pgvector</category><category>retrieval</category><category>rag</category></item><item><title>The day I sped up a slow Postgres query by 40x</title><link>https://portfolio.furkanturkoglu.workers.dev/en/blog/speeding-up-slow-postgres-query-40x/</link><guid isPermaLink="true">https://portfolio.furkanturkoglu.workers.dev/en/blog/speeding-up-slow-postgres-query-40x/</guid><description>Tracing high buffer reads in EXPLAIN ANALYZE and replacing a misordered index with a partial composite index to cut latency from 480ms to 11ms.</description><pubDate>Fri, 14 Aug 2026 00:00:00 GMT</pubDate><category>postgresql</category><category>sql</category><category>performance</category></item></channel></rss>