TechPrufer
Get started
Qdrant
Qdrant logoQdrant · Feature trail

Filtered Vector Search

54 posts · Feb 2023 – Sep 2026

What this trail covers

Qdrant continues to enhance its filtered vector search capabilities by introducing and benchmarking the ACORN algorithm alongside its existing Filterable HNSW approach. This post details the performance trade-offs between these methods, particularly for scenarios involving single and double keyword filters, and highlights the effectiveness of the query planner in selecting the optimal strategy (ACORN, Filterable HNSW, or payload index fallback) based on filter selectivity and query budget. ACORN now supports hybrid search with keyword matching and structured filters, enabling precise job matching for large datasets. Himalayas uses this to combine semantic similarity with filters for location, time zone, and employment type, achieving high recall and low latency.

All Qdrant trails
  1. How Lucy Built a Source-of-Truth Retrieval Layer for 2.8M+ SEC and DART Filings with Qdrant - QdrantThis post details Lucy's implementation of a source-of-truth retrieval layer for 2.8M+ SEC and DART filings using Qdrant. It highlights the use of Qdrant's payload filtering for precise search constrained by metadata (company, form type, date range, document ID), hybrid search (dense + BM25 with RRF) for combining semantic and lexical matching, and TurboQuant (4-bit for SEC, 2-bit for DART) for significant memory footprint reduction with minimal impact on retrieval quality. The architecture involves a router to identify the target filing, followed by parallel retrieval lanes (narrow, broad, XBRL, companion) that execute hybrid queries against Qdrant. The system achieved 94.7% Recall@10 on FinanceBench and a 0.160 gain in RAGAS Answer Accuracy compared to general web search context. Qdrant averaged 32ms per search with p90 of 82ms.
  2. How Himalayas Built Semantic Job Matching for 100,000+ Remote Listings with Qdrant - QdrantThis post details how Himalayas uses Qdrant's hybrid search and payload filtering to build a semantic job matching system for over 100,000 listings. It highlights the trade-offs of keyword search (Algolia, Elasticsearch) versus semantic search with Qdrant, emphasizing the need for both semantic similarity and precise filters (location, time zone, employment type) within a single query. The architecture combines semantic and keyword retrieval with payload filters, with Qdrant handling retrieval and filtering, and the Himalayas application handling final ranking and presentation. The post also mentions recent improvements that increased recall by 19 percentage points and cut irrelevant results by 4x.
  3. How to Tune Vector Search Without Guessing - QdrantThis post provides a detailed guide on systematically tuning Qdrant's vector search parameters, moving beyond guesswork. It introduces five linked articles that address specific tuning challenges: foundational collection settings, candidate depth for retrieval, hybrid search fusion methods (RRF vs. DBSF), the necessity and cost-effectiveness of rerankers, and performance implications of quantization when collections exceed available RAM. The post emphasizes the importance of measurable outcomes and provides a protocol for diagnosing and fixing common search quality issues, including a method for building a labeled query set for effective evaluation.
  4. Filtered Vector Search: What ACORN Fixes, and What Fixes ACORN - QdrantThis post introduces and benchmarks the ACORN-1 algorithm as a query-time enhancement for filtered vector search in Qdrant. It compares ACORN-1 against Qdrant's existing Filterable HNSW and a plain HNSW graph, analyzing performance across various filter selectivities and query budgets (`hnsw_ef`). Key findings include ACORN-1's effectiveness on fields with oversized value groups where Filterable HNSW adds no edges, and its ability to repair traversal in strict intersections. The post also details the query planner's role in dynamically selecting between ACORN, Filterable HNSW, and payload index lookups to achieve optimal recall and latency. Benchmarks demonstrate ACORN-1's trade-offs, showing higher latency than Filterable HNSW when edges are present, but outperforming it when they are not. The analysis provides concrete metrics on recall, latency, and filter checks for different configurations.
  5. Pre-Filtering vs Post-Filtering (and Why Qdrant Does Neither) - QdrantThis post introduces and details the ACORN-1 algorithm as a query-time repair mechanism for filtered vector search, complementing the index-time optimizations of Filterable HNSW. It explains how ACORN enables traversal through neighbors that fail the filter, addressing cases where Filterable HNSW's extra edges are insufficient (e.g., values shared by too many points, AND filters over broad values). The post also highlights Qdrant's query planner, which dynamically selects the optimal routing strategy (Filterable HNSW, ACORN, payload index, or full scan) based on estimated filter selectivity, moving beyond static pre-filtering or post-filtering architectures.
  7. Lessons From Building E-Commerce Search on Qdrant - QdrantThis post details building an e-commerce search system on Qdrant, focusing on practical implementation strategies. Key contributions include: demonstrating hybrid retrieval with dense and BM25 sparse vectors fused in a single query; advocating for filtering within the query for efficiency and accurate facet counts; emphasizing embedding relevant product fields over using larger models; proposing quantization in RAM without rescoring for performance gains; shifting personalization from retrieval to ranking for better control; implementing merchandising as weighted formulas in a single rescore pass; and introducing a multi-faceted evaluation approach using an LLM judge, recall@10, intrusion, and persona overlap metrics. It also identifies a specific gotcha related to query fusion and sharding.
  8. How Sunny Health Built an AI Healthcare Concierge with Qdrant - QdrantThis post details how Sunny Health migrated their provider matching workload from Postgres and fuzzy matching to Qdrant. They leveraged Qdrant's hybrid search capabilities to combine hard filters (network status, license, language) with semantic similarity and geo-ranking in a single query. This resulted in up to a 57% reduction in retrieval latency for complex queries and improved accuracy for their AI healthcare concierge. The flexible JSON payload model was crucial for handling deeply nested provider data. Future plans include using Qdrant for insurance carrier matching, semantic specialty mapping, and as long-term memory for their agent.
  9. Building real-time multimodal similarity search in Flipkart Trust & Safety with Qdrant - QdrantFlipkart implemented a real-time multimodal similarity search system using Qdrant for their Trust & Safety team. This system addresses limitations of prior batch-based methods (HBase with LSH) that took up to 9 hours for similarity searches. Qdrant was chosen for its deployment flexibility, efficient HNSW indexing, and support for high-dimensional embeddings. The new system reduced detection time to under 1 minute, integrating with the Java gRPC SDK and Prometheus metrics.
  10. How Cosmos delivered editorial-grade visual search with Qdrant - QdrantThis post details how Cosmos implemented editorial-grade visual search using Qdrant Cloud. Key technical contributions include: using named vectors to store multiple embedding types (CLIP, CNN, pHash, color) in a single collection for multimodal retrieval, implementing exact color search using CIELAB vectors and a re-ranking mechanism based on color presence, and developing an application-side fusion model for hybrid search to balance relevance, engagement, and aesthetics. This resulted in sub-second search latency and significant improvements in element processing pipelines.
  11. Qdrant 1.16 - Tiered Multitenancy & Disk-Efficient Vector Search - QdrantIntroduces the ACORN-1 algorithm as a new search option for filtered vector search. ACORN explores second-hop neighbors in the HNSW graph when direct neighbors are filtered out, improving search accuracy at the expense of performance, especially when multiple low-selectivity filters are applied. This feature can be enabled on a per-query basis via the `acorn` parameter and does not require index-time changes. Benchmarks show a significant improvement in accuracy (97.20% vs 53.34%) at a moderate latency increase (13.86 ms vs 1.25 ms) when using multiple low-selectivity filters.
  12. How OpenTable Reinvented Restaurant Discovery with Qdrant - QdrantThis post details OpenTable's use of Qdrant for their Concierge AI dining assistant. It highlights Qdrant's strengths in handling sparse embeddings for scenarios requiring fine-grained filtering down to a single restaurant out of over 60,000. The article emphasizes Qdrant's predictable performance for combined metadata and review filtering, meeting latency budgets, and the ease of deployment with Qdrant Cloud. It also notes the operational stability of Qdrant in production, allowing the OpenTable team to focus on model performance and user experience.
  13. How Alhena AI unified its AI stack and improved ecommerce conversions with Qdrant - QdrantAlhena AI migrated from FAISS and Pinecone to Qdrant Cloud, unifying their vector search infrastructure. They leveraged Qdrant's hybrid search (dense and sparse embeddings), metadata filtering, and boosting capabilities to improve e-commerce conversion rates and agent response times. The post highlights the benefits of Qdrant's multitenancy for scaling and Qdrant Cloud for simplified global deployment and feature adoption.
  14. How GoodData turbocharged AI analytics with Qdrant - QdrantThis post details GoodData's successful migration to Qdrant for their AI analytics platform, moving from an unsustainable LLM context loading approach to a RAG strategy. It highlights Qdrant's role in enabling near-real-time embedding updates and fast semantic search, contributing to AI assistant response times of 5-10 seconds. The technical details include the use of Qdrant's Helm chart for Kubernetes deployment and the performance metrics achieved (embedding updates in seconds, search results in 100ms).
  15. How FAZ unlocked 75 years of journalism with Qdrant - QdrantThis post details FAZ's implementation of a semantic search engine using Qdrant and Azure OpenAI for their 75-year journalism archive. It highlights the use of Qdrant for handling schema-less metadata payloads (over 60 fields) associated with embedded article paragraphs, enabling filtering by publication date, author, section, and article type. The system supports real-time updates and deletions, and achieves sub-second performance at scale (14M+ vectors) with quantization. FAZ also retrieves adjacent context vectors for answer generation. The post outlines plans to evolve towards a hybrid search architecture combining dense vectors, sparse vectors (BM25-like), and structured metadata filtering, with GPT-4 for query interpretation.
  16. How Lawme Scaled AI Legal Assistants and Significantly Cut Costs with Qdrant - QdrantThis case study details Lawme's migration from PGVector to Qdrant to scale their AI legal assistants. Key technical contributions include the successful implementation of Qdrant's binary quantization for efficient retrieval from tens of millions of legal vectors, and the utilization of metadata filtering with Qdrant's HNSW index to quickly filter queries by jurisdiction or case type. The post highlights the benefits of Qdrant's flexible deployment options, enabling Lawme to meet strict data residency and compliance requirements for legal clients.
  17. LegalTech Builder's Guide: Navigating Strategic Decisions with Vector Search - QdrantThis post details the application of Qdrant's Filterable HNSW and hybrid search capabilities within the LegalTech domain. It specifically highlights how pre-filtering improves speed and cost-efficiency for large legal datasets. The integration of Minicoil for hybrid search and ColBERT for token-level reranking is discussed, along with the use of score boosting to prioritize legal logic in search rankings. Performance optimizations like GPU indexing and vector quantization are mentioned in the context of scaling LegalTech applications. The post also touches upon deployment flexibility and enterprise-grade security features relevant to the legal industry.
  18. How Tripadvisor Drives 2 to 3x More Revenue with Qdrant-Powered AI - QdrantThis post details Tripadvisor's use of Qdrant to power their AI Trip Planner, which has resulted in a 2-3x revenue increase from engaged users. It highlights the creation of a user graph using Qdrant for personalized travel recommendations and conversational search, emphasizing Qdrant's suitability for real-time, unstructured data in AI applications.
  19. Precision at Scale: How Aracor Accelerated Legal Due Diligence with Hybrid Vector Search - QdrantThis post details how Aracor leveraged Qdrant's hybrid and filtered vector search capabilities to accelerate legal due diligence. It highlights the use of Qdrant for precise retrieval and accurate citations within massive legal document repositories, enabling automated summaries, comparison reports, and signature verifications. The post emphasizes Qdrant's advantages in scalability, hybrid search (combining semantic search with structured filters), and multitenancy, leading to significant improvements in workflow speed and accuracy for Aracor's clients.
  20. How Garden Scaled Patent Intelligence with Qdrant - QdrantThis post details how Garden, a patent intelligence startup, successfully scaled their platform by migrating to Qdrant Cloud. Key technical contributions include: - **Scalability:** Handling hundreds of millions of vectors and terabytes of patent data, a significant increase from previous solutions. - **Performance:** Achieving sub-100ms p95 query latency, meeting strict product requirements. - **Cost Efficiency:** Reducing cost per stored GB by approximately 10x compared to previous managed services. - **Migration Strategy:** A weekend migration from Google Cloud Storage to Qdrant Cloud with minimal code changes due to API mirroring of open-source conventions. - **Technical Features Leveraged:** Scalar quantization (8-bit) for balancing RAM and disk usage, and the core filterable HNSW capability which was a key differentiator for Garden's use case of complex metadata filtering on patent data.
  21. How Pariti Doubled Its Fill Rate with Qdrant - QdrantThis post details how Pariti integrated Qdrant into their talent marketplace to improve candidate matching. They used Qdrant Cloud for zero-ops deployment and an intuitive Python SDK. The integration involved encoding job requirements, embedding candidate experiences, and comparing vectors. This resulted in millisecond-level search results, reducing candidate vetting time from 4 minutes to 1 minute and increasing the fill rate from 20% to 48%. Pariti is considering Qdrant's GPU-assisted indexing and vector quantization for future scaling.
  22. Beyond Multimodal Vectors: Hotel Search With Superlinked and Qdrant - QdrantThis post introduces a multimodal hybrid search approach by combining Superlinked's specialized vector spaces (Text, Number, Logarithmic Scale) with Qdrant's filtering and weighted nearest neighbor search capabilities. It demonstrates how to build a hotel search demo that handles complex natural language queries by translating them into weighted vector operations and hard filters across different data domains. The post highlights the benefits of this approach over single-modal vector search and traditional hybrid search, particularly for scenarios involving diverse data types and nuanced user preferences.
  23. Qdrant 1.14 - Reranking Support & Extensive Resource Optimizations - QdrantIntroduced a Score-Boosting Reranker that allows combining vector similarity with custom logic based on payload data (e.g., prioritizing titles, boosting recent results via datetime, factoring in geographical proximity using gauss_decay). Implemented Incremental HNSW Indexing to extend existing graphs instead of rebuilding them. Optimized batch query processing by splitting large batches into chunks for concurrent execution across threads. Improved resource utilization during segment optimization by saturating CPU and disk IO more effectively. Reduced memory usage in immutable segments.
  24. Pathwork Optimizes Life Insurance Underwriting with Precision Vector Search - QdrantThis post details how Pathwork used Qdrant's filtered vector search capabilities, including scalar quantization, hybrid search, and advanced filter utilization, to significantly improve the accuracy and performance of their life insurance underwriting system. They achieved a reduction in Mean Squared Error (MSE) from 3.5 to 1.8 and decreased query latency from 9 seconds to 2 seconds. The post also highlights Qdrant's scalability features, such as sharding and replicas, which Pathwork utilized to handle a 50% month-over-month usage growth.
  25. How Lyzr Supercharged AI Agent Performance with Qdrant - QdrantThis post details how Lyzr migrated from Weaviate and Pinecone to Qdrant to improve the performance and scalability of their AI agents. It quantifies the performance gains, including a >90% reduction in query latency (from 300-500ms to 20-50ms P99), 2x faster indexing, and a 30% reduction in infrastructure costs. The benchmarks show Qdrant handling over 1,000 queries per minute and sustaining throughput of more than 250 queries per second with over 100 concurrent agents, demonstrating significant improvements in query throughput, indexing performance, and resource efficiency compared to their previous solutions. The post also includes case studies from NTT Data and NPD, highlighting Qdrant's role in improving retrieval accuracy and low-latency retrieval for AI applications.
  26. Building a Facial Recognition System with Qdrant - QdrantThis post details the implementation of a facial recognition system (Twin Celebrity app) using Qdrant. It describes the architecture involving FaceNet for embeddings, MTCNN for alignment, ZenML for orchestration, and Streamlit for the UI. The process involves an offline phase for generating and storing celebrity embeddings in Qdrant, and an online phase for querying user selfies against these embeddings. It also covers deployment strategies and the use of Qdrant's visualization tools for quality testing. Scalability considerations and potential future enhancements are discussed.
  27. Optimizing ColPali for Retrieval at Scale, 13x Faster Results - QdrantIntroduced a two-stage retrieval process for ColPali to address scaling challenges with visually rich PDFs. This involved pooling (mean and max) to reduce the number of vectors per page from 1030 to 38 for initial retrieval, followed by reranking with original high-resolution embeddings. Achieved a 13x speed improvement with mean pooling maintaining high NDCG@20 (0.952) and Recall@20 (0.917).
  28. Empowering QA.tech’s Testing Agents with Real-Time Precision and Scale - QdrantThis post details how QA.tech uses Qdrant for their AI testing agents, migrating from pgvector due to scalability issues. They leverage Qdrant's real-time analysis capabilities, batch operations to reduce network overhead, and efficient CPU utilization for embedding processing. The ability to store multiple embeddings per data point is highlighted as a key feature for their diverse use cases, enabling high-velocity lookups and supporting complex agent reasoning.
  29. Advanced Retrieval with ColPali & Qdrant Vector Database - QdrantThis post introduces ColPali, a multimodal retrieval approach that uses Vision Language Models (VLMs) to generate multi-vector embeddings directly from document images, capturing visual and textual content simultaneously. It details how ColPali works by processing document images as grids of patches, transforming them into contextualized vectors, and using a late interaction strategy with MaxSim similarity for query processing. The post also highlights the optimization benefits of Binary Quantization for ColPali embeddings when used with Qdrant, demonstrating reduced search times and memory usage compared to Scalar Quantization, and provides a practical example of using ColPali with Qdrant for searching visually rich documents.
  30. Qdrant 1.12 - Distance Matrix, Facet Counting & On-Disk Indexing - QdrantQdrant 1.12 introduces the Distance Matrix API for calculating pairwise vector distances, enabling clustering and dimensionality reduction use cases. It also adds a Facet API for counting and aggregating unique values in payload fields, similar to SQL's GROUP BY. New on-disk support for text and geo indexes reduces memory footprint by storing index data on disk, with performance considerations for cold query latency and memory vs. disk trade-offs. The Distance Matrix API is also integrated into the Web UI's Graph Exploration Tool.
  31. Nyris & Qdrant: How Vectors are the Future of Visual Search - QdrantThis post details Nyris's adoption of Qdrant for their visual product search solution. It highlights how Nyris transitioned from traditional SIFT features and early CNN embeddings to Qdrant's dedicated vector search engine. Key benefits discussed include Qdrant's accuracy, speed, insert performance, cost-effectiveness, data sovereignty, and specialized features like quantization. The integration enables Nyris to perform precise visual searches by combining image embeddings with metadata filters, moving towards a unified vector representation for multimodal data.
  32. Kern AI & Qdrant: Precision AI Solutions for Finance and Insurance - QdrantThis post details Kern AI's use of Qdrant for financial and insurance AI solutions, specifically highlighting the application of multi-vector storage and hybrid search/filters for RAG chatbots. It emphasizes achieving low hallucination rates (<1%) and reducing customer service response times. The post also mentions Qdrant's interactive Discord community and its open-source nature as key decision factors.
  33. Qdrant 1.10 - Universal Query, Built-in IDF & ColBERT Support - QdrantQdrant 1.10 introduces a Universal Query API that consolidates all search APIs (nearest, fusion, discover, context, recommend, order_by) into a single endpoint, simplifying hybrid search and other complex query combinations. It adds built-in IDF calculation for sparse vectors, removing the need for client-side computation and enabling streaming updates. Native support for late interaction ColBERT is now available via the Query API for multivector search. Sparse vector compression is also introduced. New Rust client, S3 snapshot storage, and an Issues API are also included.
  34. Visua and Qdrant: Vector Search in Computer Vision - QdrantVISUA implemented Qdrant for computer vision quality control and anomaly detection, achieving 40x faster query processing and 10x scalability. They utilized Qdrant's hybrid query capabilities, support for multiple vectors per record, and quantization. The post highlights Qdrant's advantages over competitors like Weaviate and Pinecone in handling hybrid queries and performance for VISUA's specific use cases.
  35. Advancements and Challenges in RAG Systems - Syed Asad | Vector Space Talks - QdrantThis post features a discussion about the application of RAG systems, including visual RAGs, and highlights Qdrant's seamless performance in handling complex multimedia data retrieval, emphasizing its accuracy and ease of use in such advanced applications. The speaker specifically mentions using Qdrant for a family tree app that uses video interactions with deceased loved ones, and praises Qdrant's search capabilities for this visual RAG use case.
  36. Building Search/RAG for an OpenAPI spec - Nick Khami | Vector Space Talks - QdrantThis post demonstrates building search and RAG over an OpenAPI specification using Qdrant and Trieve. It details how Qdrant's group-based system can be used to organize API endpoints by tags, enabling more structured search results. The technical implementation involves parsing the OpenAPI JSON, creating groups in Qdrant for each tag, and then inserting individual API paths as points within their respective groups. The post also briefly mentions Tantivy as a complementary search engine.
  37. VirtualBrain: Best RAG to unleash the real power of AI - Guillaume Marquis | Vector Space Talks - QdrantThis post highlights VirtualBrain's use of Qdrant as a core component of their AI assistant for knowledge workers, emphasizing RAG capabilities. It details how Qdrant's scalability, speed, low latency, and open-source nature were critical factors in VirtualBrain's decision to use it for handling large datasets (terabytes) and performing complex tasks like audits and RFPs. The post also touches on the importance of fast ingestion and semantic search to meet user expectations for quick responses, and the use of Qdrant for document retrieval and AI-assisted tasks.
  38. Production-scale RAG for Real-Time News Distillation - Robert Caulk | Vector Space Talks - QdrantThis post details the application of Qdrant in a production-scale RAG system for real-time news distillation. It highlights Qdrant's role in enabling real-time indexing and semantic similarity searches, crucial for processing over a million news articles daily. The author emphasizes Qdrant's contribution to modularity and performance in AI pipelines, contrasting it with less flexible all-in-one solutions. The post also touches upon the importance of embedding quality for effective semantic search and the benefits of Qdrant's evolving features, such as sparse vector support, in enabling such products.
  39. Vector Search for Content-Based Video Recommendation - Gladys and Samuel from Dailymotion - QdrantThis post details Dailymotion's implementation of Qdrant for content-based video recommendations. It covers their approach to generating video embeddings from metadata, audio, and transcripts, and how these embeddings are used with Qdrant to address the cold-start problem for new and niche videos. The post highlights the use of a streaming pipeline for real-time updates and the positive impact on click-through rates for low-signal content.
  40. Dailymotion's Journey to Crafting the Ultimate Content-Driven Video Recommendation Engine with Qdrant Vector Search - QdrantDailymotion implemented a content-based video recommendation engine using Qdrant. They process video metadata (title, tags, description, transcript) into MUSE embeddings and use Qdrant for fast and accurate similarity search. This resulted in a significant reduction in recommendation processing time (from hours to minutes) and a 3x increase in interaction and CTR for recommended videos, particularly for fresh and low-signal content. Qdrant's capabilities in handling high-dimensional data, scalability, and metadata filtering were crucial for this implementation.
  41. Qdrant vs Pinecone: Vector Databases for AI Apps - QdrantThis post provides a comparative analysis of Qdrant and Pinecone, focusing on their vector database capabilities. For Qdrant, it details its advanced similarity search metrics, payload indexing and filtering features (including keyword matching, full-text filtering, numerical ranges, nested object filters, and geo filtering), and its support for hybrid search with sparse vectors. It also touches upon Qdrant's Rust implementation, scaling, multitenancy, quantization, and deployment flexibility.
  42. What is Vector Similarity? Understanding its Role in AI Applications. - QdrantThis post provides a foundational explanation of vector similarity, its calculation methods (Dot Product, Cosine Similarity, Euclidean Distance, Manhattan Distance), and its significance in AI applications. It also outlines the challenges of scalability, efficiency, and security in vector similarity search and briefly mentions Qdrant's role in addressing these challenges.
  43. Dust and Qdrant: Using AI to Unlock Company Knowledge and Drive Employee Productivity - QdrantThis post details how Dust integrated Qdrant into their AI assistant platform, focusing on retrieval augmented generation (RAG). Key technical aspects include Dust's two-step approach: starting with Qdrant Cloud for quick setup and then scaling and optimizing by leveraging Qdrant's features like quantization (Scalar Quantization) and MMAP payload threshold configuration. The post highlights the resulting cost reduction (2x) and ability to maintain low latency and high accuracy at scale. Observability features of Qdrant were also noted as beneficial for scaling.
  44. The Bitter Lesson of Retrieval in Generative Language Model Workflows - Mikko Lehtimäki | Vector Space Talks - QdrantThis post discusses the 'bitter lesson' in AI, emphasizing the power of data and compute over handcrafted methods. It delves into retrieval-augmented generation (RAG) workflows, specifically how the Yokot AI platform handles diverse data inputs and the importance of re-ranking for improving LLM output quality. The discussion touches on data flexibility, document uploading, web scraping, and the evolution of information retrieval methods.
  45. Qdrant Updated Benchmarks 2024 - QdrantThis post introduces updated benchmarks for Qdrant, comparing its performance against other vector search engines. Key changes include applying suggestions for more efficient engine operation, resulting in up to four times improvement in certain cases. A new dataset of 1 million OpenAI embeddings for RAG applications has been added. The benchmark methodology now separates latency and requests-per-second (RPS) scenarios, simulating 1 or 100 parallel readers. The post reiterates Qdrant's commitment to open-source, accessible benchmarks, and focuses on production-ready vector databases over indexing libraries.
  46. Optimizing an Open Source Vector Database with Andrey Vasnetsov - QdrantThis post details Qdrant's strategy for in-place filtering during graph traversal in vector search, which optimizes precision and exactness when scaling to billions of vectors. It introduces techniques like using subgraphs with overlapping intervals and geo hash regions for enhanced precision and connectivity. The post also discusses controlling search precision and speed by adjusting the beam size in HNSW indices and highlights the challenges of maintaining graph connectivity during filtering. Additionally, it touches upon vector compression methods.
  47. Vector Search Complexities: Insights from Projects in Image Search and RAG - Noé Achache | Vector Space Talks - QdrantThis post discusses the application of vector search in image search and RAG projects. It highlights the effectiveness of the Dino V2 model for image representation, noting its ability to understand objects and patterns without fine-tuning. The post also touches on challenges in data deduplication for real estate listings and the use of multi-language embedding models for document retrieval in RAG. It mentions the importance of efficient storage and retrieval, including using memory for CI integration and the complexities of metadata filtering in vector databases.
  48. How to Superpower Your Semantic Search Using a Vector Database Vector Space Talks - QdrantMalt successfully transitioned to Qdrant for their freelancer matching application, significantly reducing latency from 10 seconds to 1 second. This involved adopting a retriever-ranker architecture, utilizing multilingual transformer-based encoder models for high-fidelity embeddings, and leveraging Qdrant's capabilities for geospatial filtering and scaling.
  49. Building LLM Powered Applications in Production - Hamza Farooq | Vector Space Talks - QdrantThis post discusses building LLM-powered applications in production, focusing on complex search, conversational search, and personalization. It highlights challenges in LLM service downtime and strategies for mitigation, as well as the unique approach of Traversal AI in creating unified platforms for LLM integration and domain-specific search. The importance of user experience (UX) in GenAI product development is also emphasized.
  50. Building a High-Performance Entity Matching Solution with Qdrant - Rishabh Bhardwaj | Vector Space Talks - QdrantThis post details the use of Qdrant for high-performance entity matching, specifically for hotel data. It highlights the challenges of data inconsistency, duplication, and real-time processing across multiple languages. The solution leverages Qdrant's HNSW algorithm for efficient vector indexing, demonstrating significant improvements in speed and recall compared to initial PostgreSQL experiments with Pgvector. The post also touches upon the importance of geofiltering for accurate matching and the experimentation with different embedding models (Mini LM) and HNSW parameters (M and efConstruct) to optimize performance.
  51. When music just doesn't match our vibe, can AI help? - Filip Makraduli | Vector Space Talks - QdrantThis post details the application of Qdrant's vector search capabilities for a music recommendation system. It describes a novel approach to generating song embeddings using natural language descriptions queried from LLMs, and then encoding these descriptions with sentence transformers. The core of the recommendation system relies on Qdrant for performing cosine similarity search on these embeddings to find songs matching user-defined moods or vibes. The post highlights the use of Qdrant's vector indexing and search for this specific use case.
  52. Chat with a codebase using Qdrant and N8N - QdrantThis post demonstrates the integration of Qdrant with n8n to build a "chat with a codebase" service. It details two n8n workflows: one for ingesting GitHub repository data into Qdrant, and another for querying this data to provide chat functionality. The ingestion workflow utilizes Qdrant's Vector Store Insert node, GitHub Document Loader, OpenAI Embeddings, and Recursive Character Text Splitter. The chat workflow uses Qdrant's Vector Store Retrieve node, Retrieval Q&A Chain, OpenAI Embeddings, and OpenAI Chat Model. It provides specific node configurations and prerequisites for setting up the system.
  53. Pienso & Qdrant: Future Proofing Generative AI for Enterprise-Level Customers - QdrantThis post details how Qdrant's vector computation capabilities, specifically its efficient nearest neighbor search and distributed deployment mode, enhance Large Language Models (LLMs) for enterprise customers like Pienso. It highlights Qdrant's ability to provide dynamic long-term storage, preserve extensive context windows, and enable scalable, cost-efficient high vector computation. Technical details include storage efficiency benchmarks (128 million documents in 20.4GB storage), the use of Rust for performance, memmap storage for fast access, and write-ahead logging for data safety.
  54. Powering Bloop semantic code search - QdrantThis post details how bloop uses Qdrant for its semantic code search engine. It explains the problem of semantic search for code, the benefits of using embeddings and vector search, and how Qdrant provides fast, accurate, and reliable vector similarity search for large codebases. The post emphasizes Qdrant's performance and ease of integration via its Rust client.