aws

PartyRock from AWS is a new application from AWS that allows anybody to generate a Generative AI application in minutes. It uses Amazon Bedrock to not only run the text and image generation, but also to generate the apps in the first place. How does it work?

Large Language Models and Generative AI are big topics in 2023, and whilst using them as a service is great (and sensible), I wondered if it was possible to run a LLM completely within a Lambda Function.

AWS DeepRacer is a small, self-driving car providing a hands-on introduction to reinforcement learning. Training can be costly and time-consuming. I'm experimenting with training on cost-effective EC2 instances and sharing the knowledge as the competitive nature of the league often restricts shared information.

Automated Machine Learning (AutoML) is a field of study and technology aimed at automating the end-to-end process of designing and building machine learning models. I explore the basics and key concepts of AutoML using Amazon's SageMaker Autopilot.

Recently Amazon OpenSearch Serverless was announced that allows the use of the same service but will automatically scale and where you pay for the resources consumed, this is split over the cost of Compute and cost of Storage.

Infrastructure as Code (IaC) is an important part of Cloud Infrastructure, by utilizing flexible cloud resources, preferably serverless, then it opens up the possibility of a pipeline to create a full Developer environment per Branch, and then destroy it again after the Pull Request is merged.