InnerVault

RAG Pipeline taste

Indexedprompts.chat (CC0 prompt collection), f

# RAG Pipeline taste of an engineer who ships retrieval systems that stay accurate - Build an evaluation set of real user questions with expected sources before tuning chunking, embeddings or prompts. Confidence: 0.95…

Open the source (opens github.com) https://github.com/f/prompts.chat

What it is

Prompts from f.

The problem it solves and who it suits haven't been reviewed by InnerVault. The description above is the source's own.

The prompt

# RAG Pipeline taste of an engineer who ships retrieval systems that stay accurate
- Build an evaluation set of real user questions with expected sources before tuning chunking, embeddings or prompts. Confidence: 0.95
- Measure retrieval and generation separately: recall@k and MRR for retrieval, faithfulness and answer quality for generation. Confidence: 0.90
- Chunk by document structure such as headings, sections and paragraphs before falling back to fixed token windows. Confidence: 0.85
- Keep chunks around 300-800 tokens with modest overlap, then tune against the eval set instead of guessing. Confidence: 0.75
- Store rich metadata with every chunk: source URL, title, section path, updated date, permissions and content hash. Confidence: 0.90
- Prepend the document title and section heading to each chunk's embedded text so isolated chunks keep their context. Confidence: 0.80
- Use hybrid search, combining BM25 keyword search with dense vectors, because names, codes and IDs embed poorly. Confidence: 0.85
- Add a cross-encoder reranker over the top 30-50 candidates before passing the best 5-10 chunks to the model. Confidence: 0.80
- Apply access-control filters inside the retrieval query, never by filtering results after the model has seen them. Confidence: 0.95
- Version the embedding model and re-embed the whole corpus when it changes; never mix vectors from different models in one index. Confidence: 0.90
- Make ingestion incremental and idempotent using content hashes, so re-running it only touches changed documents. Confidence: 0.85
- Start with Postgres and pgvector for small to medium corpora before adding a dedicated vector database. Confidence: 0.70
- Ask the model to cite chunk IDs for every claim and render those citations as links back to the source. Confidence: 0.90
- Instruct the model to say it doesn't know when retrieved context is insufficient, and test that behavior explicitly. Confidence: 0.90
- Rewrite conversational follow-up questions into standalone queries before retrieval. Confidence: 0.80
- Treat retrieved text as untrusted data and guard against prompt injection hidden inside documents. Confidence: 0.85
- Log the query, retrieved chunk IDs, scores and final answer for every request so failures can be replayed. Confidence: 0.85
- Prefer fixing the corpus, chunking or metadata over adding longer and longer system prompts. Confidence: 0.75
- Cache embeddings for repeated queries and set a token budget for context rather than stuffing in everything retrieved. Confidence: 0.70
- Run the eval suite in CI on every change to prompts, chunking, models or retrieval settings. Confidence: 0.85
- Reach for fine-tuning only after measurements show retrieval is no longer the bottleneck. Confidence: 0.65

Written by the prompts.chat (CC0 prompt collection) community (f). Not tested by InnerVault.

How to get it

Setup steps: not applicable.

Related

  1. a lot of people even professional said a software product or platfrom especially which has complex rules which was fully coded by AI by a vibe coder who deosnt understand what is the arcture concept deployed, read and…

    Indexed hussen.sepa@gmail.com · via prompts.chat (CC0 prompt collection) CC0-1.0

  2. code-hacker

    Prompts

    Code Hacker - VS Code Custom Agent + Web + TUI

    Stale hackerlibs · via GitHub 55 ★MITUpdated 20 days ago

    Source for code-hacker (opens github.com) git clone https://github.com/hackerlibs/code-hacker.git
  3. builder prompts which can be sent to builders like cursor, bolt.diy, bolt.new, trae, project idx, Claude code, lovable.dev, llmacoder, geminicoder, windsurf, deepseek-artifact, cline, aider etc

    Stale KingLeoJr · via GitHub 278 ★MITUpdated 2 months ago

    Source for awesome-vibe-coding-prompt-code-templates (opens github.com) git clone https://github.com/KingLeoJr/awesome-vibe-coding-prompt-code-templates.git
  4. Imagine you are an experienced Ethereum developer tasked with creating a smart contract for a blockchain messenger. The objective is to save messages on the blockchain, making them readable (public) to everyone…

    Indexed ameya-2003 · via prompts.chat (CC0 prompt collection) 177.9k ★CC0-1.0

  5. Manim Code

    Prompts

    Your task to create a manim code that will explain the chain rule in easy way

    Indexed industry5designers@gmail.com · via prompts.chat (CC0 prompt collection) CC0-1.0

  6. I’m tired of using Claude Code to build my code because of tokens limits can Ollama build code scripts agentic workflow?

    Indexed BuiltByPhil · via prompts.chat (CC0 prompt collection) CC0-1.0