An introduction to Amazon Bedrock, generative AI on AWS

This article gives an easy explanation of Bedrock, the generative AI service on AWS. We will also publish an article on Amazon SageMaker, another AWS generative AI service, so please read that one too.

Since generative AI became widely available in 2022, it has attracted a great deal of attention. More and more businesses are using generative AI, including the generative AI services on AWS.

This article is for you if you want to

• learn about generative AI on AWS
• learn about Bedrock
• choose the generative AI that fits your company

What is Bedrock?

Bedrock is a service that helps a lot when developing generative AI applications like ChatGPT.

For example, imagine building an application that automatically writes blog posts like this one. What steps would the development take? It would probably look like this.

1. Planning: who will use it, what kind of service it will be, and how much it will cost
2. Investigation: how to build it and how to teach it language
3. Research: whether it can output well after loading large amounts of data
4. Development: building it based on the research results
5. Completion: release

Of these steps, steps 2 to 4 take the most time and effort. These alone can often take months or even years.

Bedrock helps with this hard work. It offers a wide range of foundation models (FMs) from Amazon and leading startups. A foundation model is one for which steps 2 and 3 above are already done.

You can choose an FM that fits your use case and customize it, for example by adding other functions you need.

You could compare it to a travel agency. When you tell a travel agency where you want to go, such as Europe or the United States, it offers various package tours. You choose the one that fits your needs, and you may add a visit to a restaurant you have been curious about.

In the same way, Bedrock is like a marketplace that offers packages combining AWS resources.

Benefits of Bedrock

1. A wide range of models

Bedrock offers many FMs, so you can choose one for each use case, such as text generation, text summarization, or chat systems. In other words, you can choose the best FM for your needs.

2. Highly customizable

You can customize the FM you choose. This lets developers specialize AI models for a specific industry or domain and use them in application development.

3. Works well with AWS

Because Bedrock is part of AWS, development and deployment on AWS are easy, and it integrates smoothly with existing AWS services.

That was an introduction to Bedrock. We hope it gave you an idea of its benefits and how it can be used.

If you have any questions, as an official AWS Partner, we are happy to answer them and make proposals. Please feel free to contact us.