4 ms·
On my own time and dime, I came up with a way to glue together a thin-layer open-source LLM and open-source Meilisearch. After initial testing, it works exceedi
by dserban 3y ago
On my own time and dime, I came up with a way to glue together a thin-layer open-source LLM and open-source Meilisearch. After initial testing, it works exceedingly well for ecommerce / retail / online shopping, for search, discovery, product recommendation and deep intent recognition.
Now I have to figure out how to turn this into a monetized product, in the face of legal threats from my employer.
(I happen to work for one of the entrenched-retrograde "IBMs" of the information retrieval industry.)
- Kerollmops 3y agoThat looks pretty cool. Would you mind explaining a little bit how you did that into more detail? Like, you ask Meilisearch some documents (keyword search) and then ask the LLMs score of those and refine the list a bit?
- dserban 3y agoIt's a two-step prompt-injection process. In the first step, I wrap the raw typed-in input inside of a dynamically-generated prompt template that takes some context into account (for example, the specifics of the product that the person is currently looking at, or the recent history thereof). The prompt generates a list of Meilisearch keyword search queries. I issue all queries and harvest the results. Scoring and ranking happens in the second (also prompt-injection powered) step. So the answer to your question is pretty much yes.