TAGGED TOPIC
#Rag
Explore all engineering breakdowns, benchmarks, and tutorials tagged with #Rag.
11Articles Found
Showing 1–11 of 11 ArticlesPage 1 of 1
Hybrid Search + Reranking for Production RAG: Vector, BM25, RRF, Cross-Encoder Patterns 2026AI
2026-10-2112 min read
Hybrid Search + Reranking for Production RAG: Vector, BM25, RRF, Cross-Encoder Patterns 2026
Practical production patterns for hybrid retrieval: parallel BM25 + vector search, RRF fusion, and cross-encoder reranking with measured latency and tuning guidance.
Asutosh Sidhya
Practical Architecture for Cross-Domain Hallucination Evaluation: Lessons from MultiLLM-ChatbotAI
2026-10-1011 min read
Practical Architecture for Cross-Domain Hallucination Evaluation: Lessons from MultiLLM-Chatbot
Scalable evaluation architecture for detecting LLM hallucinations across domains and metrics; synthesizes MultiLLM-Chatbot methods into production assessment patterns.
Asutosh Sidhya
Production-grade LLM Hallucination Detection: Benchmarks, Metrics, and RAG PipelineAI
2026-10-0811 min read
Production-grade LLM Hallucination Detection: Benchmarks, Metrics, and RAG Pipeline
A production blueprint for detecting and mitigating LLM hallucinations using multiple benchmarks, detectors, and a reproducible RAG evaluation pipeline with metrics.
Asutosh Sidhya
Advanced Hybrid RAG: HyDE, BM25+Vector Fusion, and Cross-Encoder RerankingAI
2026-10-0610 min read
Advanced Hybrid RAG: HyDE, BM25+Vector Fusion, and Cross-Encoder Reranking
Engineer's guide to Hybrid RAG (2026): HyDE, BM25+vector fusion, cross-encoder reranking, chunking, GraphRAG patterns, benchmarks, and production code.
Asutosh Sidhya
Engineering an LLM Hallucination Detection Pipeline: Benchmarks, Metrics, and Production PatternsAI
2026-10-0110 min read
Engineering an LLM Hallucination Detection Pipeline: Benchmarks, Metrics, and Production Patterns
Design and deploy a reliable hallucination detection pipeline for LLMs using modern benchmarks, AUROC evaluation, RAG-aware detectors, and production-grade monitoring and remediation.
Asutosh Sidhya
Haystack 2.0 Architectural Guide: Modular Production RAG PipelinesFrameworks
2026-08-044 min read
Haystack 2.0 Architectural Guide: Modular Production RAG Pipelines
Master deepset Haystack 2.0 — covering component-based directed graph pipelines, InMemoryDocumentStore, OpenAIGenerator, conditional routers, and evaluation nodes.
Asutosh Sidhya
LlamaIndex Architectural Guide: Data Ingestion, Indexing, and Query EnginesFrameworks
2026-08-024 min read
LlamaIndex Architectural Guide: Data Ingestion, Indexing, and Query Engines
A comprehensive developer handbook for LlamaIndex — covering VectorStoreIndex, Document Readers, node parsing, sub-question query engines, and agentic RAG workflows.
Asutosh Sidhya
LangChain v0.3+ Architectural Guide: LCEL Runnables, Tools, and RetrievalFrameworks
2026-07-224 min read
LangChain v0.3+ Architectural Guide: LCEL Runnables, Tools, and Retrieval
Master modern LangChain v0.3+ architecture — covering LangChain Expression Language (LCEL), streaming runnables, document loaders, vector retrievers, and structured output parsing.
Asutosh Sidhya
Building a Multi-Agent AI Framework (Part 2/5): Hybrid Memory & Context PruningAI
2026-03-2411 min read
Building a Multi-Agent AI Framework (Part 2/5): Hybrid Memory & Context Pruning
Step 2 in our 5-part masterclass: Implementing short-term token-sliding windows, vector memory indexing, and dynamic context pruning.
Asutosh Sidhya
Next-Gen Agentic RAG: Hybrid Search, GraphRAG, and Self-CorrectionAI
2026-02-256 min read
Next-Gen Agentic RAG: Hybrid Search, GraphRAG, and Self-Correction
Architecting enterprise retrieval systems using LanceDB, hybrid dense-sparse vector indexing, GraphRAG knowledge graphs, and dynamic re-ranking loops.
Asutosh Sidhya
Understanding Retrieval-Augmented Generation (RAG)AI
2026-02-118 min read
Understanding Retrieval-Augmented Generation (RAG)
A practical deep dive into building production-grade RAG pipelines — covering document chunking strategies, embedding models, hybrid BM25 + dense retrieval, context window management, and the evaluation metrics that actually matter.
Asutosh Sidhya