explore | Grilled Cheese

New arXiv paper evaluates LLM retrieval robustness in RAG setups, asking whether retrieval always helps, if more documents improve results, and whether document order matters, using a 1,891-sample benchmark across five datasets.

#LLMs #RAG #AIResearch
https://arxiv.org/abs/2505.21870

"Beyond the Prompt: How RAG Gives LLMs Access to Your Own Data on AWS" by Sarah Enaile

#rag #llm #aws-builders

🤖 How do you know if a RAG system is giving reliable answers?

🔍 RAG Evaluation measures retrieval quality, context relevance, faithfulness, answer correctness & hallucination risks.

Read the article: www.linkedin.com/pulse/what-r...

#RAG #RAGEvaluation #GenerativeAI #AI #LLM

I’ve been building a Python-based multi-agent framework with LangGraph, RAG, FastAPI, MCP & tool-based workflows. I’d really appreciate feedback from developers working on AI agents: what would you improve or add?
github.com/himanshukuma...

#AI #AIAgents #LangGraph #RAG

The best embedding model isn’t necessarily the one that performs best on a public benchmark.

It’s the one that retrieves the right answers from your own data.

That’s why I built Truvec:

→ Bring your RAG data
→ Benchmark embedding models
→ Compare retrieval quality
→ Find what actually works

#RAG

How do you turn a local RAG app in C# into an AI Agent?
In Part 1, I use Microsoft Agent Framework to expose Qdrant search as a tool and let the model decide when retrieval is needed.
www.ottorinobruni.com/from-local-r...
#dotnet #csharp #aiagent #rag #ollama #qdrant

Un millón de #Embeddings de 1.024 dimensiones en float32: ~4 GB. Casi nunca hace falta tanto.

Matryoshka a 256 + binaria: ~32 MB, 128 veces menos. Y con rescoring recuperas buena parte de la calidad.

Mide contra el original: el ahorro que vale es el que no se nota en tu #RAG.

"Building a RAG Pipeline on AWS with Amazon Bedrock, pgvector, and FastAPI" by Vighnesh Sankpal

#amazon-bedrock #amazon-bedrock-agents #vector-database #amazon-s3-vectors #rag

"RAG Using AWS: Connecting AI With Your Own Data" by Jatin Adlak

#rag #aws-codebuild #agentic-ai #ai #aws-builders

【ざっくり理解】ベクトルDBは新しい技術なのか？ ─ 1975年からRAGまで、古いものと新しいものを分けて見る

#ベクトル検索 #べクトルDB #rag #LLM #歴史

#BizzSummitES is a wrap!

It was a great experience to be in Madrid and share the stage with my dear friend @fabiofranzini.com , talking about how we can make enterprise knowledge truly usable by #AI.
#BizzSummitES #MicrosoftCopilot #CopilotStudio #AzureAISearch #RAG #AI #MVPBuzz

Systems Thinking for Agentic AI

An LLM is powerful , but an LLM alone is not an AI system.

Now available on
📘 Leanpub
leanpub.com/staai

📗 Google Books play.google.com/store/books/...

Ediz Najim
staaicore.com

#agenticai #aiengineering #SoftwareArchitecture #LLM #RAG #MCP #SoftwareEngineering

🔧 Runs on #vLLM with Aleph Alpha's inference plugin, exposing an OpenAI-compatible API. Needs 2x A100/H100 80 GB or 1x H200/B200/B300. Best fit: on-prem #RAG over long German documents like laws, contracts and manuals

#PostgreSQL #in-process with #vectorsearch extensions directly #embedded in your #Python app. #OpenSource Apache 2.0 lic

- Similarly search: #pgvector + #pgvectorscale, high-performance storage
- #FullText search: #pg_textsearch, #BM25 and #ranking
#AI #LLM #Embeddings #Agents #RAG