We live in the digital era, and creating data becomes easier every day. According to Forbes, we generate 2.5 quintillion bytes of data each day, and this data comes from all types of sources. It can be pictures on Instagram, voice messages on Telegram, videos, text, or even a combination of all of them. This was definitely not the case when the internet was just starting. And yet, despite the obvious difference in the easiness of data creation before and now, we keep using the same search techniques. Despite having an incredible boom in data generation, we haven’t updated our search techniques that much.

Neural search is the approach to changing that. It takes advantage of the machine learning (ML) era that we live in right now and uses the latest AI research to deliver novel search techniques. However, despite this being a good technique, it presents a lot of challenges. Neural search is a whole new concept, which means the knowledge and techniques needed are also new. To effectively deploy a neural search application, the engineers in charge of it need to have a plethora of skills, from engineering to Dev-Ops, and an ML background.

This is the problem Jina AI wants to solve. Jina is an open source solution that is designed to democratize AI and neural search, making it easier for many developers to have a full end-to-end neural search application without having to have a background in ML, the cloud, and backend engineering. It’s designed with Python developers in mind to help them unlock the full potential of the latest neural search techniques. In this book, we will explore the basics of search, from traditional to neural search. With this knowledge, we will work through hands-on examples of creating a fully fledged neural application.