Understanding embeddings without the math degree

May 21, 2024 #ai #llm

Vectors, similarity, and why your search results suddenly got better.

I kept a log of every decision during this project, and reading it back a month later was humbling: the choices that aged well were the boring ones, and the clever ones all came with a footnote I hadn't written yet.

Background

I want to be honest about the trade-offs here. The approach I settled on solves the problem I had, not every problem of this shape, and there is a clear point where you should reach for something more heavy-duty.

The first version worked, which is exactly why it was hard to see how wrong it was. Everything looked fine in the demo, and everything looked fine until traffic arrived on a Tuesday afternoon and the graph started doing something I'd never seen before.

[deploy]
strategy = "rolling"
healthcheck = "/healthz"
max_unavailable = 1

How it works

What surprised me most was how small the final change was. Most of the work was understanding the problem precisely enough that the fix became obvious, and that understanding took far longer than the fix itself.

A tip for a practical recommendation.

In practice

Tools matter less than the feedback loop they enable. A fast, boring loop beat every sophisticated setup I tried, and I've kept it ever since, even as the rest of the stack changed twice over.

By the end of this post you should be able to reproduce the whole thing yourself in under an hour, including the parts I got wrong the first time — those are marked so you can skip them if you're in a hurry.

Latency at p50, because most users live there
Latency at p99, because someone always lives there
Error rate, because a fast failure is still a failure
Saturation, because averages hide the cliff

Edge cases

There is a version of this article that opens with theory. I decided against it, because the theory is easy to find and the motivation is not — so let's start with the thing that actually broke.