5 ms·You serve the embedding model in a lambda and then run something like FAISS in the backend.by bootsmann 3y agoYou serve the embedding model in a lambda and then run something like FAISS in the backend.