Production RAG: what I learned from processing 5M+ documents

(blog.abdellatif.io)

265 points | by tifa2up 7 hours ago ago

73 comments

  • bityard 4 hours ago

    I must be missing something, this says it can be self-hosted. But the first page of the self-hosting docs say you need accounts with no less than 6 (!) other third-party hosted services.

    We have very different ideas about the meaning of self-hosted.

    • RobertDeNiro 3 hours ago

      That was my observation as well. To be fair their business is to sell a hosted version, they’re under no obligation to release a truly self hosted version.

    • dgfitz 3 hours ago

      I’ve never worked in such a space where the deployed environment had unfettered internet access, no access at all actually.

      I’ve probably missed a huge wave of programming technology because of this, and I’ve figured out a way to make it work for a consistent paycheck over these past 20 years.

      I’m also not a great example, I think I’ve watched 7 whole hours of YouTube videos ever, and those were all for car repair help.

      I shy away from tech that needs to be online/connected/whatever.

    • goodev 4 hours ago

      I consider this to be good open source and I'm a happy user of their OSS offering. Want no hosted dependencies? Then go write it all in Rust.

      • icemanx 34 minutes ago

        that's a stupid take and shows lack of engineering experience

  • mediaman 6 hours ago

    The point about synthetic query generation is good. We found users had very poor queries, so we initially had the LLM generate synthetic queries. But then we found that the results could vary widely based on the specific synthetic query it generated, so we had it create three variants (all in one LLM call, so that you can prompt it to generate a wide variety, instead of getting three very similar ones back), do parallel search, and then use reciprocal rank fusion to combine the list into a set of broadly strong performers. For the searches we use hybrid dense + sparse bm25, since dense doesn't work well for technical words.

    This, combined with a subsequent reranker, basically eliminated any of our issues on search.

    • siva7 5 hours ago

      Boy, that should not be the concern of the end user (developer) but those implementing RAG solutions as a service at Amazon, Microsoft, Openai and so on.

      • pamelafox 3 hours ago

        At Microsoft, that's all baked into Azure AI Search - hybrid search does BM25, vector search, and re-ranking, just with setting booleans to true. It also has a new Agentic retrieval feature that does the query rewriting and parallel search execution.

        Disclosure: I work at MS and help maintain our most popular open-source RAG template, so I follow the best practices closely: https://github.com/Azure-Samples/azure-search-openai-demo/

        So few developers realize that you need more than just vector search, so I still spend many of my talks emphasizing the FULL retrieval stack for RAG. It's also possible to do it on top of other DBs like Postgres, but takes more effort.

        • catmanjan 2 hours ago

          I'd love to work with Azure search but because copilot with external items has been made so cheap it's hard to justify...

        • osigurdson an hour ago

          Are you using Elasticsearch behind the scenes?

          • pamelafox an hour ago

            I believe that Azure AI Search currently uses lucene for BM25, hnswlib for vector search, and the Bing re-ranking model for semantic ranking. (So, no, it does not, though features are similar)

        • alansaber 2 hours ago

          That is concerning given that pure vector search is terrible outside of abstractions

          • pamelafox 2 hours ago

            I know :( But I think vector DBs and vector search got so hyped that people thought you could switch entirely over to them. Lots of APIs and frameworks also used "vector store" as the shorthand for "retrieval data source", which didn't help.

            That's why I write blog posts like https://blog.pamelafox.org/2024/06/vector-search-is-not-enou...

            • osigurdson an hour ago

              It is almost like embeddings are a technology from the olden days.

    • deepsquirrelnet 5 hours ago

      > For the searches we use hybrid dense + sparse bm25, since dense doesn't work well for technical words.

      One thing I’m always curious about is if you could simplify this and get good/better results using SPLADE. The v3 models look really good and seem to provide a good balance of semantic and lexical retrieval.

    • alansaber 2 hours ago

      Yep- that's all best practice. I want to know if we could push performance further- routing the query to different embedding models or scoring strategies, or using multiple re-rankers- still feels like the process is missing something.

    • avereveard 5 hours ago

      final tip is to also feed the interpretation of the user search to the user on the other side, so he can check if the llm understanding was correct.

  • daemonologist 6 hours ago

    I concur:

    The big LLM-based rerankers (e.g. Qwen3-reranker) are what you always wanted your cross-encoder to be, and I highly recommend giving them a try. Unfortunately they're also quite computationally expensive.

    Your metadata/tabular data often contains basic facts that a human takes for granted, but which aren't repeated in every text chunk - injecting it can help a lot in making the end model seem less clueless.

    The point about queries that don't work with simple RAG (like "summarize the most recent twenty documents") is very important to keep in mind. We made our UI very search-oriented and deemphasized the chat, to try to communicate to users that search is what's happening under the hood - the model only sees what you see.

    • agentcoops 4 hours ago

      I agree completely with your point, especially the difficulty of developing the user's mental model for what's going on with context and the need to move away from chat UX. It's interesting that there are still few public examples of non-chat UIs that make context management explicit. It's possible that the big names tried this and decided it wasn't worth it -- but from comments here it seems like everyone that has built a production RAG system has come to the opposite conclusion. I'm guessing the real reason is otherwise: likely for the consumer apps controlling context (especially for free users) and inference time is one of the main levers for cost management at scale. Private RAGs, on the other hand, are more concerned with maximizing result quality and minimizing time spent by employee on a particular problem with cost per query much less of a concern --- that's been my experience at least.

    • thethimble 5 hours ago

      I wish there was more info on the article about actual customer usage - particularly whether it improved process efficiency. It's great to focus on the technical aspects of system optimization but unless this translates to tangible business value it's all just hype.

  • js98 6 hours ago

    Similar writeup I did about 1.5 years ago for processing millions of (technical) pages for RAG. Lots has stayed the same it seems

    https://jakobs.dev/learnings-ingesting-millions-pages-rag-az...

    • winstonp 5 hours ago

      I also built a RAG system about a year back for technical search, everything seems the same!

  • hatmanstack 5 hours ago

    Not here to schlep for AWS but S3 Vectors is hands down the SOTA here. That combined with a Bedrock Knowledge Base to handle Discovery/Rebalance tasks makes for the simplest implementation on the Market.

    Once Bedrock KB backed by S3 Vectors is released from Beta it'll eat everybody's lunch.

    • arcanemachiner 5 hours ago

      Shill, not schlep.

      I'm correcting you less out of pedantry, and more because I find the correct term to be funny.

      • hatmanstack 5 hours ago

        I feel like I'm schelpin' through these comments, it's all mishigas

        • esafak 4 hours ago

          You feel like a schlemiel, perhaps?

          • hatmanstack 4 hours ago

            more a schlimazel, Charles Schultzie, Lucy's everywhere

      • latchkey an hour ago

        Especially now that if you google the word schlep, the first result is now something totally different than what you'd expect.

  • leetharris 6 hours ago

    Embedding based RAG will always just be OK at best. It is useful for little parts of a chain or tech demos, but in real life use it will always falter.

    • phillipcarter 5 hours ago

      Not necessarily? It's been the basis of one of the major ways people would query their data since 2023 on a product I worked on: https://www.honeycomb.io/blog/introducing-query-assistant

      The difference is this feature explicitly isn't designed to do a whole lot, which is still the best way to build most LLM-based products and sandwich it between non-LLM stuff.

    • underlines 5 hours ago

      rag will be pronounced differently ad again and again. it has its use cases. we moved to agentic search having rag as a tool while other retrieval strategies we added use real time search in the sources. often skipping ingested and chunked soueces. large changes next windows allow for putting almost whole documents into one request.

    • sgt 6 hours ago

      What do you recommend? Query generation?

    • esafak 6 hours ago

      Compared with what?

      • leetharris 4 hours ago

        Full text agentic retrieval. Instead of cosine similarity on vectors, parsing metadata through an agentic loop.

        To give a real world example, the way Claude Code works versus how Cursor's embedded database works.

        • lifty 3 hours ago

          How do you do that on 5 million documents?

    • charcircuit 6 hours ago

      Most of my ChatGPT queries use RAG (based on the query ChatGPT will decide if it needs to search the web) to get up to date information about the world. In reality life it's effective and it's why every large provider supports it.

  • osigurdson an hour ago

    Speaking of embedding models, OpenAIs are getting a little long in the tooth at this stage.

  • n_u 6 hours ago

    > Reranking: the highest value 5 lines of code you'll add. The chunk ranking shifted a lot. More than you'd expect. Reranking can many times make up for a bad setup if you pass in enough chunks. We found the ideal reranker set-up to be 50 chunk input -> 15 output.

    What is re-ranking in the context of RAG? Why not just show the code if it’s only 5 lines?

    • tifa2up 5 hours ago

      OP. Reranking is a specialized LLM that takes the user query, and a list of candidate results, then re-sets the order based on which ones are more relevant to the query.

      Here's sample code: https://docs.cohere.com/reference/rerank

      • yahoozoo 4 hours ago

        What is the difference between reranking versus generating text embeddings and comparing with cosine similarity?

        • derefr 3 hours ago

          My understanding:

          If you generate embeddings (of the query, and of the candidate documents) and compare them for similarity, you're essentially asking whether the documents "look like the question."

          If you get an LLM to evaluate how well each candidate document follows from the query, you're asking whether the documents "look like an answer to the question."

          An ideal candidate chunk/document from a cosine-similarity perspective, would be one that perfectly restates what the user said — whether or not that document actually helps the user. Which can be made to work, if you're e.g. indexing a knowledge base where every KB document is SEO-optimized to embed all pertinent questions a user might ask that "should lead" to that KB document. But for such documents, even matching the user's query text against a "dumb" tf-idf index will surface them. LLMs aren't gaining you any ground here. (As is evident by the fact that webpages SEO-optimized in this way could already be easily surfaced by old-school search engines if you typed such a query into them.)

          An ideal candidate chunk/document from a re-ranking LLM's perspective, would be one that an instruction-following LLM (with the whole corpus in its context) would spit out as a response, if it were prompted with the user's query. E.g. if the user asks a question that could be answered with data, a document containing that data would rank highly. And that's exactly the kind of documents we'd like "semantic search" to surface.

        • osigurdson an hour ago

          Because LLMs are a lot smarter than embeddings and basic math. Think of the vector / lexical search as the first approximation.

        • tifa2up 4 hours ago

          text similarity finds items that closely match. Reranking my select items that are less semantically "similar" but are more relevant to the query.

  • esafak 6 hours ago

    They say the chunker is the most important part, but theirs looks rudimentary: https://github.com/agentset-ai/agentset/blob/main/packages/e...

    That is, there is nothing here that one could not easily write without a library.

    • tifa2up 6 hours ago

      OP here. We've been working on agentset.ai full-time for 2 months. The product now gets you something working quite well out of the box. Better than most people with no experience in RAG (I'd say so with confidence).

      Ingestion + Agentic Search are two areas that we're focused on in the short term.

    • teraflop 6 hours ago

      I'm not sure there is a chunker in this repo. The file you linked certainly doesn't seem to perform any chunking, it just defines a data model for chunks.

      The only place I see that actually operates on chunks does so by fetching them from Redis, and AFAICT nothing in the repo actually writes to Redis, so I assume the chunker is elsewhere.

      https://github.com/agentset-ai/agentset/blob/main/packages/j...

  • pietz 5 hours ago

    I find it interesting that so many services and tools were investigated except for embedding models. I would have thought that's one of the biggest levers.

    • Trias11 5 hours ago

      they just grabbed the better one (3-large) right off the bat. 6x cost to 3-small, but it's still tiny.

      • pietz an hour ago

        But the model is like 18 months old. and recently we've seen big leaps on MTEB. Not sure how well those translate to reality, but I'm a little surpised this wasn't worth looking into.

  • whinvik 2 hours ago

    Anybody know what is meant by 'injecting relevant metadata'. Where is it injected?

  • jascha_eng 6 hours ago

    I have a RAG setup that doesn't work on documents but other data points that we use for generation (the original data is call recordings but it is heavily processed to just a few text chunks). Instead of a reranker model we do vector search and then simply ask GPT-5 in an extra call which of the results is the most relevant to the input question. Is there an advantage to actual reranker models rather than using a generic LLM?

    • tifa2up 6 hours ago

      OP here. rerankers are finetuned small models, they're cheap and very fast compared to an additional GPT-5 call.

      • jascha_eng 6 hours ago

        It's an async process in my case (custom deep research like) so speed is not that critical

    • alansaber 2 hours ago

      I think you should do both in parallel, rather than sequentially. Main reason is vector scoring could cut off something that an LLM will score as relevant

  • manishsharan 6 hours ago

    Thanks for sharing. TIL about rerankers.

    Chunking strategy is a big issue. I found acceptable results by shoving large texts to to gemini flash and have it summarize and extract chunks instead of whatever text splitter I tried. I use the method published by Anthropic https://www.anthropic.com/engineering/contextual-retrieval i.e. include full summary along with chunks for each embedding.

    I also created a tool to enable the LLM to do vector search on its own .

    I do not use Langchain or python.. I use Clojure+ LLMs' REST APIs.

    • esafak 6 hours ago

      Have you measured your latency, and how sensitive are you to it?

      • manishsharan 6 hours ago

        >> Have you measured your latency, and how sensitive are you to it?

        Not sensitive to latency at all. My users would rather have well researched answers than poor answers.

        Also, I use batch mode APIs for chunking .. it is so much cheaper.

  • bitpatch 3 hours ago

    Really solid write-up — it’s rare to see someone break down the real tradeoffs of scaling RAG beyond the toy examples. The bit about reranking and chunking actually saving more than fancy LLM tricks hits home to me.

  • 383toast 5 hours ago

    They should've tested other embedding models, there are better ones than openai's (and cheaper)

  • alexchantavy 6 hours ago

    > What moved the needle: Query Generation

    What does query generation mean in this context, it’s probably not SQL queries right?

    • andreasgl 5 hours ago

      I think they mean query expansion: https://en.wikipedia.org/wiki/Query_expansion

    • daemonologist 6 hours ago

      It's described in the remainder of the point - they use an LLM to generate additional search queries, either rephrasings of the user's query or bringing additional context from the chat history.

      • goleary 5 hours ago

        Here's an interesting read on the evolution beyond RAG: https://www.nicolasbustamante.com/p/the-rag-obituary-killed-...

        One of the key features in Claude Code is "Agentic Search" aka using (rip)grep/ls to search a codebase without any of the overhead of RAG.

        Sounds like even RAG approaches use a similar approach (Query Generation).

        • smokel 4 hours ago

          The article raises several interesting points, but I find its claim that Claude Code relies primarily on grep for code search unconvincing. It's clear that Claude Code can parse and reason about code structure, employing techniques far beyond simple regex matching. Since this assumption underpins much of the article's argument, it makes me question the overall reliability of its conclusions a bit.

          Or am I completely misunderstanding how Claude Code works?

  • dcreater 4 hours ago

    do you still use langchain/llamaindex for other agents/AI use cases?

  • inshard 2 hours ago
  • nextworddev 6 hours ago

    Exactly what kind of processing was done? Your pipeline is a function of the use case, lest you overengineer…