Articles by tag: rag

AI agent with RAG and MCP: integration of external tools

Assembling an AI agent on Python that uses RAG for documents and MCP for external APIs. Step-by-step implementation with code and architectural explanations.

Memory OS: reasoning paradox and SGR in agent memory

Analysis of Memory OS architecture: from flat RAG failure to memory graph on 106 million tokens. SGR, ConceptHypothesis, LongMemEval for developers. Study the lessons.

Graph Product Search on LLM

LLM Graph Search Architecture for Complex Queries: Negations, Structures, Brands. Pipeline, Examples, Quality Assessment. For RAG System Developers — Implement in 10–15s.

PageIndex in RAG: replacement for vector search

Analysis of PageIndex for RAG without embeddings: pros, cons, local launch on Ollama. Testing on PDF, accuracy 69% with qwen3:14b. Setup instructions for developers.

DRAG with KNEE for RAG: dynamic pruning

Learn how DRAG with KNEE improves RAG through hierarchical Qdrant trees and knee-point pruning. Adaptive search reduces noise and tokens. Implementation in Python.

RAG: smart document search for developers

Break down the implementation of Retrieval-Augmented Generation: from embeddings to hybrid search and RRF. Code, algorithms, optimizations for middle/senior. Build your own system.

AI assistant for calls: from prototype to production

Learn how a backend team without ML experience created a real-time voice assistant on RAG, BERT, Qwen 8B. Architecture, compromises, lessons for developers. Read the case.

FAQ Automation with AI: Vector Search and RAG

Explore technical methods for automating FAQ responses: embeddings, no-code bots, GenAI with RAG. Instructions for middle/senior dev, implementation examples. Speed up customer service without losses.

RAG optimization for legal tasks

Learn how hybrid RAG with semantic chunking and grounding improved metrics by 2x in Agentic RAG Legal Challenge. Benchmarks, pipelines and lessons for developers.

Graph instead of RAG for regulatory documents

Why vector search breaks on regulations: from flat chunks to graph of nodes and links. Multi-mode retrieval, terms, mandatory context. For middle/senior dev. Learn more.

Master PM tech stack: AI assistant on FastAPI

Step-by-step guide for PM: SaaS development with FastAPI, PostgreSQL, RAG. 10 stages from Telegram bot to hybrid search. Deepen the technique without simplifications.

AI-agent for technical specification verification: development, architecture, practice

Practical experience creating an AI-agent for auditing technical specifications. Learn about data collection, error classification, and hybrid RAG + decision trees architecture.