skills.rest
lagarcess
lagarcessCOMMUNITY
Explore all skills by this author →
GitHub

rag-architect

Design and optimize retrieval-augmented generation pipelines for document indexing and vector search.

1|Updated Apr 4, 2026
Search Tags:#embeddings#rag#evaluation#retrieval#chunking#reranking#vector-database
One-click install
npx skills add https://github.com/lagarcess/argus --skill rag-architect-lagarcess
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: rag-architect
Source: https://github.com/lagarcess/argus/tree/main/.agent/skills/rag-architect
Command: npx skills add https://github.com/lagarcess/argus --skill rag-architect-lagarcess

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Design, implement, and optimize end-to-end Retrieval-Augmented Generation (RAG) pipelines to reliably surface relevant knowledge from large document collections while minimizing hallucinations, latency, and cost. The skill reduces time spent on trial-and-error decisions around chunking, embedding selection, vector database choice, retrieval strategies, reranking, and evaluation so teams can deliver grounded answers with source attribution.

Core Features & Use Cases

  • Document chunking strategies: Guidance on fixed-size, sentence, paragraph, semantic, recursive, and document-aware chunking with overlap and hierarchy preservation.
  • Embedding model selection: Recommendations and trade-offs for MiniLM, mpnet, e5, bge, and OpenAI embeddings across speed, quality, and cost.
  • Vector database integration: Patterns for choosing and configuring Pinecone, Weaviate, Qdrant, Chroma, and pgvector including indexing and filtering best practices.
  • Retrieval & reranking: Dense, sparse, hybrid retrieval patterns, multi-query and HyDE techniques, and cross-encoder reranking for precision.
  • Evaluation & production patterns: Metrics (NDCG, MRR, precision@K, faithfulness), A/B testing, monitoring, caching, streaming retrieval, and safety guardrails for PII and hallucination detection.
  • Use Case: Build a customer support knowledge base that semantic-searches product docs and returns source-linked, faithful answers with automated reranking and monitoring.

Quick Start

Index the product documentation using semantic chunking, e5 or bge embeddings, Qdrant for vector search, and validate retrieval quality with NDCG and MRR.

Frequently Asked Questions about rag-architect

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
What's the best way to design a production-grade RAG pipeline?▼

Designing a production-grade RAG pipeline involves selecting semantic chunking strategies, choosing embeddings like e5 or bge, integrating vector databases such as Qdrant, and applying cross-encoder reranking to ensure accurate, grounded knowledge retrieval.

How do I choose the right embedding model for my retrieval-augmented generation system?▼

Choosing the right embedding model requires evaluating trade-offs across speed, quality, and cost for MiniLM, mpnet, e5, bge, and OpenAI embeddings to match your specific retrieval and latency requirements.

What chunking strategy should I use for document indexing in RAG?▼

Chunking strategies for document indexing range from fixed-size and paragraph methods to semantic, recursive, and document-aware chunking, preserving hierarchy and overlap to maintain context during retrieval.

How do I evaluate retrieval quality and faithfulness in a RAG pipeline?▼

Evaluate retrieval quality and faithfulness using metrics like NDCG, MRR, precision@K, and faithfulness checks, alongside A/B testing and monitoring to detect hallucinations and ensure source attribution.

Does cross-encoder reranking improve dense retrieval accuracy in vector databases?▼

Cross-encoder reranking improves dense and hybrid retrieval accuracy by reordering top vector database results from Pinecone, Weaviate, or Chroma, increasing precision for downstream generation.

How do I minimize hallucinations when building a customer support knowledge base with RAG?▼

Minimize hallucinations in a customer support RAG knowledge base by applying safety guardrails, PII detection, faithfulness evaluation, and automated reranking to return source-linked, grounded answers.

Related Skills

Explore More Software Engineering Skills→
vaaynevaayne
52

kreuzberg

Extract text, tables, metadata, and images from 91+ document formats.

Community
Advanced
llama-farmllama-farm
835

rag-skills

Audit and optimize RAG pipelines with LlamaIndex, ChromaDB, and Celery best practices.

Official
Basic
Sir-chawakornSir-chawakorn
9

build-vector-search

Configure embedding models, ANN indices, and hybrid search pipelines for vector retrieval.

Community
Advanced
peteknowsaipeteknowsai

knowledge

Store and semantically retrieve personal knowledge locally with Stoolap and Gemini embeddings.

Community
Advanced