Transcription
What is Amazon Bedrock? So, in simple lines, Amazon Bedrock is a serverless application or a service within AWS, in simple.
So now, why someone needs to go to Amazon Bedrock is now the very first thing. Going to be now, if I want to connect with any of the LLM models, especially foundational models, I need to create an account in OpenAI. I need to get the API key. I need to add a balance, or I need to go to Anthropic. I need to create a login in Anthropic. I want to get an API key from Anthropic in order to use cloud models. Now, in order to use any of the models, you need to have an account and then you need to be able to create an API key.
So now, if you are able to have Bedrock, so with the help of Bedrock, what's going to happen? You are able to access multiple different providers of your foundational models. You can see here. So AI21 Labs, the list of models which are available from AI21 Labs, from Amazon. These are the list of models. You can see I'm able to see access granted. I can request the access and I can pay for their usage. And again, if I scroll, Anthropic based models, Claude models, all the latest models, Cohere models, DeepSeek models, Meta models, Mist AI Labs, so on and so forth. Based on your location, you are able to access different models.
And not only that, for in order to develop your chatbots and all that, so you need to customize your models, you need to build RAG architecture, you want to come up with fine-tuning, you want to come up with pre-training. So all that you are able to do it within Bedrock by using AWS SageMaker. And even not only that, for example, some companies don't really believe in OpenAI and all that because of the security issues. Now, with the help of Amazon, you are able to have the proper responsible AI within Amazon Bedrock. So with the proper AI based guardrails, we are able to find it within Amazon Bedrock. That's one of the reasons most of the organizations are relying on Bedrock. And not only that, so for deployment, for managing your entire solution, if it is from AWS, that's going to be even more easy for a guy who wants to release their entire thing with the help of LLM Ops. So that's one of the reasons where they are knocking the door of Amazon Bedrock. It's very simple.
Now, see this. Now, if I want to connect with Amazon Bedrock, I want to have an AWS CLI, and this is how I'm connecting Amazon Bedrock. I'm giving my profile credentials, AWS CLI credentials, model ID, and the model means like arguments I'm just providing. That's it. As simple as that. Like connecting your OpenAI, you are connecting with Bedrock, and you are able to access.