9 ms·
Production RAG: what I learned from processing 5M+ documents
- manishsharan 1y agoThanks 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 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 1y agoHave you measured your latency, and how sensitive are you to it?
- manishsharan 1y 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.
- crassT 1y agoI made a startup, https://tokencrush.ai/ https://tokencrush.ai/, to do just this. I've struggled to find a target market though. Would you mind sharing what your use case is? It would really help give me some direction.
- jascha_eng 1y agoI 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 1y agoOP here. rerankers are finetuned small models, they're cheap and very fast compared to an additional GPT-5 call.
- jascha_eng 1y agoIt's an async process in my case (custom deep research like) so speed is not that critical
- alansaber 1y agoI 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
- esafak 1y agoThey say the chunker is the most important part, but theirs looks rudimentary: https://github.com/agentset-ai/agentset/blob/main/packages/engine/src/chunk.ts 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 1y agoOP 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 1y agoI'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/jobs/src/tasks/process-document.ts#L185 https://github.com/agentset-ai/agentset/blob/main/packages/j...
- alexchantavy 1y ago> What moved the needle: Query Generation What does query generation mean in this context, it’s probably not SQL queries right?
- daemonologist 1y agoIt'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 1y agoHere's an interesting read on the evolution beyond RAG: https://www.nicolasbustamante.com/p/the-rag-obituary-killed-by-agents 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 1y agoThe 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?
- andreasgl 1y agoI think they mean query expansion: https://en.wikipedia.org/wiki/Query_expansion https://en.wikipedia.org/wiki/Query_expansion
- js98 1y agoSimilar 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-azure/ https://jakobs.dev/learnings-ingesting-millions-pages-rag-az...
- winstonp 1y agoI also built a RAG system about a year back for technical search, everything seems the same!
- nextworddev 1y agoExactly what kind of processing was done? Your pipeline is a function of the use case, lest you overengineer…
- daemonologist 1y agoI 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.
- thethimble 1y agoI 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.
- agentcoops 1y agoI 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.
- leetharris 1y agoEmbedding 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.
- sgt 1y agoWhat do you recommend? Query generation?
- esafak 1y agoCompared with what?
- leetharris 1y agoFull 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 1y agoHow do you do that on 5 million documents?
- leetharris 1y agoPeople are usually not querying across 5 million documents in a single scope. If you want something as simple as "suggest similar tweets" or something across millions of things then embeddings still work. But if you want something like "compare the documents across these three projects" then you would use full text metadata extraction. Keywords, summaries, table of contents, etc to determine data about each document and each chunk.
- charcircuit 1y agoMost 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.
- mediaman 1y agoThe 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.
- avereveard 1y agofinal 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.
- deepsquirrelnet 1y 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.
- siva7 1y agoBoy, 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 1y agoAt 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/ 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.
- n_u 1y 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 1y agoOP. 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 https://docs.cohere.com/reference/rerank
- yahoozoo 1y agoWhat is the difference between reranking versus generating text embeddings and comparing with cosine similarity?
- tifa2up 1y agotext similarity finds items that closely match. Reranking my select items that are less semantically "similar" but are more relevant to the query.
- derefr 1y agoMy 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.
- 383toast 1y agoThey should've tested other embedding models, there are better ones than openai's (and cheaper)
- prettyblocks 1y agoWhich do you suggest?
- roze_sha 1y agohttps://huggingface.co/spaces/mteb/leaderboard https://huggingface.co/spaces/mteb/leaderboard
- 383toast 1y agoyep
- remz14 1y agoYou should use RTEB instead. See here for why: https://huggingface.co/blog/rteb https://huggingface.co/blog/rteb Here is that leaderboard https://huggingface.co/spaces/mteb/leaderboard?benchmark_name=RTEB%28beta%29 https://huggingface.co/spaces/mteb/leaderboard?benchmark_nam... Voyage-3-large seems like SOTA right now
- leftnode 1y agoThe Qwen3 600M and 4B embedding models are near state of the art and aren't too computationally intensive.
- hatmanstack 1y agoNot 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 1y agoShill, not schlep. I'm correcting you less out of pedantry, and more because I find the correct term to be funny.
- hatmanstack 1y agoI feel like I'm schelpin' through these comments, it's all mishigas
- esafak 1y agoYou feel like a schlemiel, perhaps?
- hatmanstack 1y agomore a schlimazel, Charles Schultzie, Lucy's everywhere
- latchkey 1y agoEspecially now that if you google the word schlep, the first result is now something totally different than what you'd expect.
- cipherself 1y agoS3 Vectors is hands down the SOTA here SOTA for what? Isn't it just a vector store?
- DSingularity 1y ago
- pietz 1y agoI 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 1y agothey just grabbed the better one (3-large) right off the bat. 6x cost to 3-small, but it's still tiny.
- pietz 1y agoBut 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.
- jokethrowaway 1y agoi'd go with qwen embedding 3, gemini embeddings or something from mixedbread
- bityard 1y agoI 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.
- goodev 1y agoI 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 1y agothat's a stupid take and shows lack of engineering experience
- dgfitz 1y agoI’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.
- RobertDeNiro 1y agoThat 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.
- nl 1y agoYou can self-host their code. I don't think there is any official definition of "self hosted" that this violates. For example - if a "self hosted" service supports off-site backups is it self hosted or just well designed?
- 1y ago
- dcreater 1y agodo you still use langchain/llamaindex for other agents/AI use cases?
- _pvzn 1y agoReally 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.
- whinvik 1y agoAnybody know what is meant by 'injecting relevant metadata'. Where is it injected?
- tifa2up 1y agoYou typically add a lot of metadata with each chunk text to be able to filter it, and do to include in the citations. Injecting metadata means that you see what metadata adds helpful context to the LLM, and when you pass the results to the LLM you pass them in a format like this: Title: ... Author: ... Text: ... for each chunk, instead of just passing the text
- inshard 1y agoNice app bro https://usul.ai/chat/VgnzXjlRdljIDMBVCfqiy https://usul.ai/chat/VgnzXjlRdljIDMBVCfqiy
- osigurdson 1y agoSpeaking of embedding models, OpenAIs are getting a little long in the tooth at this stage.
- mattfrommars 1y agoGreat read. But how do people land opportunities to work on exciting project as the author did? I've been trying to get into legal tech in LLM space but I've been unsuccessful. Anyone here successfully transitioned into legal space? My gut always been legal to the space where LLM can really be useful, the first one is in programming.
- torrmal 1y agowe have been trying to make it so that people dont have to reinvent the wheel, over and over and over again, and have a very straight forward all batteries included that can scale to many millions of documents, combining the best of RAG with traditional search and parametric search, https://docs.mindsdb.com/mindsdb_sql/knowledge_bases/overview https://docs.mindsdb.com/mindsdb_sql/knowledge_bases/overvie... Would love your feedback.
- jweewee 1y agoDoes anyone know how to do versioning for embeddings? Let’s say I want to update/upsert my data and deliver v6 of domain data instead of v1 or filter for data within a specified date range. I am thinking of exploring context prepending to chunks.
- bob_theslob646 1y agoThis is a great question
- meander_water 1y agoYour vector store should let you store the original text as well as metadata, where you can store the version. For e.g. turbopuffer lets you filter on attributes https://turbopuffer.com/docs/query#filtering https://turbopuffer.com/docs/query#filtering
- captainregex 1y agoHow much of a hit would you take on quality if you moved the processing local? have you experimented with it? don’t think llamaindex has local sadly
- tifa2up 1y agoQuite a decent hit. Local models don't perform very well in long contexts. We're planning to support a local-only offline set-up for people to host w/o additional dependencies
- liqilin1567 1y ago> Chunking Strategy: this takes a lot of effort, you'll probably be spending most of your time on it Could you share more about chunking strategies you used?
- watat83 1y ago[dead]
- swyx 1y ago> LLM: GPT 4.1 -> GPT 5 -> GPT 4.1, covered by Azure credits whats this roundtrip? also the chronology of the LLM (4.1) doesnt match the rest of the stack (text-embedding-large-3), feels weird
- tifa2up 1y agoOP. We migrated to GPT-5 when it came out but found that it performs worse than 4.1 when you pass lots of context (up to 100K tokens in some cases). We found that it: a) has worse instruction following; doesn't follow the system prompt b) produces very long answers which resulted in a bad ux c) has 125K context window so extreme cases resulted in an error Again, these were only observed in RAG when you pass lots of chunks, GPT-5 is probably a better model for other taks.
- swyx 1y agolove the share, ty
- urbandw311er 1y agoTo somebody thinking of building or paying for such a RAG system, would a workable solution be: * Upload documents via API into a Google Workspace folder * Use some sort of Google AI search API on those documents in that folder …placing documents for different customers into different folders. Or the Azure equivalent whatever that is.
- max002 1y agoGreat post, gonna be super useful for me :)
- pietz 1y agoMy biggest RAG learning is to use agentic RAG. (Sorry for buzzword dropping) - Classic RAG: `User -> Search -> LLM -> User` - Agentic RAG: `User <-> LLM <-> Search` Essentially instead of having a fixed loop, you provide the search as a tool to the LLM, which does three things: - The LLM can search multiple times - The LLM can adjust the search query - The LLM can use multiple tools The combination of these three things has solved a majority of classic RAG problems. It improves user queries, it can map abbreviations, it can correct bad results on its own, you can also let it list directories and load files directly.
- googamooga 1y agoI fully support this approach! When I first started experimenting—rather naively—with using tool-enabled LLMs to generate documents (such as reports or ADRs) from the extensive knowledge base in Confluence, I built a few tools to help the LLM search Confluence using CQL (Confluence Query Language) and store the retrieved pages in a dedicated folder. The LLM could then search within that folder with simple filesystem tools and pull entire files into its context as needed. The results were quite good, as long as the context didn’t become overloaded. However, when I later tried to switch to a 'Classic RAG' setup, the output quality dropped significantly and I refrained from switching.
- jokethrowaway 1y agoyes but the assistant often doesn't search when it should and very rarely does multiple search rounds (both on gpt5 or on claude sonnet 4.5, weaker models are even worse at tool calling)
- pietz 11mo agoCannot confirm this. Both sound like prompting issues. - Depends on your use case to let the model understand when and when not to use tools - gpt-5 s VERY persistent and often searches more than 10 times in a single run depending on the results. We're using pydantic AI where the entire Agent loop is taken care of by the framework. Highly recommend.
- badlogic 1y agoI run a few production RAG systems, some as old as end of 2023 and arrived at the same conclusions. Query expansions and non-naive chunking give the biggest bang for the bug, with chunking being the most resource intensive task, if the input data is chunk (pun intended).
- logon1915 1y agoHow did you evaluate this system? Did you make use of open source frameworks like Ragas or had develop custom metrics if so, How?