Rinha de Backend 2026 - Part 2

Mon, 11 May 2026 10:20:00 +0000

The main principle to solve this edition is Vector Search, so before entering this topic, I decided to understand what embeddings are, which is fundamental to understand and implement a Vector Search.

Embeddings are a way of representing objects like images, audio and texts as points in a vector space. The locations of those points are helpful to find similar objects.

Example

For better understanding, we will use examples of fruits represented as vectors. Each dimension of the vector represents an attribute of the fruit; we will use: sweetness and color. Each dimension value must be between 0 and 1, the green color is equal to 0, and the red color is equal to 1.

Rinha de Backend 2026 - Part 1

Mon, 11 May 2026 00:02:00 +0000

Rinha de Backend in a code challenge where you compete with other developers to build a backend under CPU, memory, and architecture constraints. The theme of the year is fraud detection using vector search.

For more information, please visit the Rinha de Backend Repository .

This project already had previous editions. In this one, I will document the process of building my solution to this Code Challenge. This edition has a base principle that I’m unfamiliar with: Vector Search. The next posts will be reporting my progress and what I’m learning in the process.

Vector Search on ermesonqueiroz

Rinha de Backend 2026 - Part 2

Example

Rinha de Backend 2026 - Part 1