# Overview of AWS GenAI Services | Bedrock | AgentCore | Q | Quick | Kiro

## Метаданные

- **Канал:** AWS with Chetan
- **YouTube:** https://www.youtube.com/watch?v=xp4HqvJoFXc
- **Дата:** 08.05.2026
- **Длительность:** 9:42
- **Просмотры:** 156

## Описание

In this short video, let me give you high level overview of AWS GenAI services landscape by breaking down the services into 3 layers - Infrastructure layer , Platform Layer and Application Layer.

If you are interested to get started with AWS checkout this course: https://www.udemy.com/course/aws-certified-cloud-practitioner-clf-c02-latest/?referralCode=E4E802AB0D47D58AC70A

If you are interested to become AWS Solutions Architect certified then checkout this course: https://www.udemy.com/course/new-aws-certified-solutions-architect-associate-saa-c03/?referralCode=03B514139A84A6F72055

## Содержание

### [0:00](https://www.youtube.com/watch?v=xp4HqvJoFXc) Segment 1 (00:00 - 05:00)

Hi, I'm Chetan, and in this video, let me walk you through AWS GenAI services. And here we are going to talk about Amazon Bedrock, Agent Core, Amazon Q, Kiro, and some of the agents that AWS has launched. Now, this is a short video to give you just the overview of these all AWS GenAI services and not to really dive deeper into individual AWS services. For that, I will release separate videos in this AI series. All right. So, the way we will look at AWS GenAI services is from the lens of who is using that services and what's the purpose of using a particular GenAI service. And in that context, we can divide AWS GenAI services in three layers. And the first layer is infrastructure layer, where you can use AWS hardware to build your own foundational models and the large language models. So, basically, here you are using AWS hardware and all the latest generation chips, networks to build your own models. So, that's the infrastructure layer. Now, the second layer is the platform layer, where you don't care how the model is built, but you just want to use the FM and the LLM from all the leading model provider out there. For example, if you want to use cloud model from Anthropic or Lama from Meta, and likewise so many other models, then you can directly use it over an API call. And further, if you want to build your own agents, so for that as well, you need a platform. So, this middle layer is all about using the AWS platform to use the FMs or to build your agents. However, if you don't even want to do that and you don't worry about the underlying model, and if you want to directly use GenAI powered applications, then you can go to the top layer where AWS provides GNAI integrated into AWS services, and also it provides ready-to-use agents for different tasks. Right? So, which means depending on your requirement, you can choose to operate in one of the three layers, or you can also use all three layers at the same time. All right. So, now let's talk about the infrastructure layer. So, here AWS provides you all the leading and most modern infrastructure to build the FMs and LLMs from scratch. So, here you get all the power from AWS. For example, the strongest of the GPUs and CPUs. You get the custom-built chips like Inferentia and Trainium. Further, you get Graviton arm-based processor, which AWS has built. So, basically using all this hardware, you can train and build your own large language models. And in order to manage your data, in order to evaluate your models, and everything, you can further use Amazon SageMaker along with this infrastructure. So, this layer is for GNAI experts and practitioners, and this provides you maximum flexibility and the customization options. Now, while this is good, understand that building the FM and the LLM takes lot of computing resources. And that's where it is really, really expensive to build these models from scratch. And that's where only big enterprises like Meta, Amazon, Google, Anthropic, Open AI, all these can afford to build these models from scratch. So, this is really expensive thing. Now, for most of the other companies who are consumer of the AI, probably middle and top layer makes more sense. So, as you go to the middle layer, there is Amazon Bedrock service and agent core service, which allows you to use this model and to build your own agent. Now, if you talk about Amazon Bedrock, it's a platform and it allows you to use all the leading FMs and LLMs from all these companies over an API call. Now, apart from that, Bedrock provides a lot more functionalities. For example, it allows you to build your own agents as well. Further, if you want to fine-tune your models, then you can do that with Amazon Bedrock. And on top of that, it provides a lot of guardrails. And as you might know, as you're using AI, you need to make sure that it is safe to use the AI services. For example, if you are a finance company, then you have to make sure that AI is not leaking any internal information as somebody interacting with your model. So, in order to safeguard your AI application, you have to apply and use these guardrails. So, basically, Amazon Bedrock platform offers everything that you need to build your AI-powered application using the leading FM and LLM models out there. All right. So, that's about using the FM and LLMs. But as you know, we are now moving towards the agentic AI. And that's where you might need a platform which allows you to build your own agent. And when we say agent, you can

### [5:00](https://www.youtube.com/watch?v=xp4HqvJoFXc&t=300s) Segment 2 (05:00 - 09:00)

use different agentic frameworks. Like, for example, you can use LangChain, LangGraph, CrewAI, Strands Agent. So, basically, Agent Core is a framework agnostic and it allows you to use all these frameworks to build your agent. And on top of that, it provides agent runtime, as in how the agent will actually run. Then it provides the security, it provides the memory. And on top of that, it also provides agent gateway, which allows your agent to talk to other agents or other systems and tools over MCP or A2A protocol. Right? So, that's the middle layer. Now, third, which is a top layer, is the application layer and there you will see a lot of new tools coming every day. And at the center there is Amazon Q service which has lot of variations. For example, Q provides developer version. So, it assists developer to write the code. Otherwise, you can also use Amazon Q with your businesses. For example, you can feed in all your business data to the Q and you can ask questions about your business. So, that's in business version and on top of that, Q service is now integrated with many AWS services. So, as of today, Q is integrated with Amazon Q site which is a BI tool, then with Amazon Connect which is kind of call center service and then with AWS supply chain. And I'm sure in the near future, Q will be integrated with many other such AWS services. Right. So, that's about Amazon Q. Now, on top of that, AWS is now also building ready-to-use agents. For example, as of today, that is April 2026, AWS has released couple of ready-to-use agents like DevOps agent which basically provides everything that you need to know and act upon as a DevOps engineer. So, it looks at any incidences that might affect your production system. It tries to find the root cause of the issue. It creates a run book to fix the issue and if required, it can also go and fix that issue. So, it is a very powerful companion for your DevOps teams. Now, other than that, for the security as well, AWS has built security agent which stays along in your complete journey from development to the production. Now, apart from these two agents, there is also one more very powerful agent and that's Q robot. Now, you can simply install the Q robot as an CLI and you can ask Q robot to write the code or debug your problem and it will do everything. And I'm not joking. Since I installed Q robot into my workstation, I interact with this all the time. So, I recommend you that try Amazon Q and see how it helps you in your day-to-day work. All right, so these are all the agents that AWS currently has and I'm sure over the time there will be new agents that AWS will launch. Now, further the last part, AWS has recently launched a productivity tool and that's Amazon Q or Amazon Q Suite, which basically improves your productivity by connecting all the productivity tools together. For example, your email, your Slack, then any of your other productivity tools that you use. So, AWS claims that Amazon Q really boost your productivity. Now, personally I haven't tried this tool yet, but as I will try, I will provide more insights as in how it works. So, as of now that's it really about the GenAI application layer and as you can see this slide has already gone very busy and this is going to expand further in the near future. All right, so I hope this slide gives you that one-stop view of all AWS GenAI services and you find this useful. Now, before we stop for this video, just one thing that I would recommend is that do try Amazon Q agent because you can also use it for free up to certain credits and you will definitely find value in that. Apart from that, if you are into the GenAI space, then I think you should also learn Agent Core because the future is all about building the agentic system and you need a platform which allows you to build these secure agents and for that, I think Agent Core will be very important to learn. So, that's it really about this video and in the future I'm going to come up with many, many AWS generative AI related videos. So, stay tuned and please subscribe and like to this channel. Thank you and I'll see you into the next video.

---
*Источник: https://ekstraktznaniy.ru/video/50052*