What AWS Services Should I Use for My Generative AI Project?

Before diving into services, make sure you have your account security sorted out—set up Multi-Factor Authentication (MFA) and billing alerts. Trust me, you wouldn’t want to deal with unexpected charges later!

Absolutely! When starting with AWS, focus on understanding account security and billing. Many newcomers overlook these aspects, and it can lead to complications down the road.

If you're open to alternatives, I've heard good things about GCP too!

Use Bedrock for LLMs along with S3 for both vector databases and storage. Then create your MVP with API Gateway and ECS to manage app integration and caching.

For your AI needs, I recommend using AWS Bedrock. It's great for LLMs. You should also check out the new S3 Vector DB, which could be useful. For image and video storage, S3 is the go-to. If you need to cache your application, DynamoDB is a good choice as a serverless key-value database, or you might consider RDS for traditional databases like PostgreSQL or MySQL.

For your setup, consider using App Runner or ECS with Fargate for the application. Stick with S3 for storage, use the S3 Vector for your database, and definitely check out Amazon Bedrock for LLMs. Also, don’t forget to apply for AWS startup credits; they can really help!