---
title: "Discussion on \"How I Cut AI Search Latency by 35% Using Parallel Fan-Out Architecture\" | Hashnode"
description: "Discussion on \"How I Cut AI Search Latency by 35% Using Parallel Fan-Out Architecture\". A model swap did not fix our slow AI search, even though the model takes ~4 seconds to respond. It only worked when i changed the architecture. Here is what I measured, what I learned from Google, and"
url: "https://hashnode.com/posts/how-i-cut-ai-search-latency-by-35-using-parallel-fan-out-architecture/6ac4c7cfebfe797868373014"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [How I Cut AI Search Latency by 35% Using Parallel Fan-Out Architecture](https://blog.desmondezoojile.com/how-i-cut-ai-search-latency-by-35-using-parallel-fan-out-architecture?utm_source=hashnode&utm_medium=feed)

A model swap did not fix our slow AI search, even though the model takes ~4 seconds to respond. It only worked when i changed the architecture. Here is what I measured, what I learned from Google, and

Oct 6 · 14 min read · blog.desmondezoojile.com  
Read the article

## Written by

[Desmond Ezo-Ojile](https://hashnode.com/@Dezy)

@Dezy

  
[View profile](https://hashnode.com/@Dezy)

Turning ideas into reality...

I'm a full stack software Engineer with over a decade of combined experience building apps in web 2 and web 3.

## Responses

Markdown supported · ⌘↵ to post

No responses yet.
