Got the #LLM semantic to structural matching working. Testing with local #AI#olama . 4b models take 2 mins and not consistent and a little 1B is more accurate and takes 5 seconds! Need to investigate a rag vector filter upstream.
Implementing my own #dos measures for public forms. Honeypots and rate limiting. Any “free” or cheap custom form services will get you price wise if you actually expect any volume on your site.
#AWS apprunner does not play nice with godaddy dns for custom domain validations. Their new two step breaks with underscores in the DNS on godaddy. Forced to use route53.
Using #AWS#kiro to dockerize my apps with specific requirements. It creates scripts, docs, and mult env configs, and deployment YAML. I don’t want to be a docker expert. It’s a breeze.
It energizes me being a #fullstack dev. #react and #ui one minute and #agentic#ai agents in python the next. Throw in C#, serverless cloud services, APIs, infrastructure, and deployment packages. I can build anything.
So many take the easy route because of speed or simplicity. Like #vibecoding push deploy. This leads to vendor lock in and high costs when you scale. Write in a way to make the switch between platforms or services an easier lift. #coding#solopreneur
Trying #AWS#kiro for #AI coding some new projects. Using requirements, design, and implementation task documents. Dials in unit test, integration tests, and more.. spec driven development.
Google just released a python library for data extraction!
LangExtract is a python library that extracts structured information from unstructured text documents with precise source grounding and interactive visualization.
100% Open Source
Working with #AWS CDK to build stacks using #lambda functions from Images. Can't escape the custom #docker images for Lambdas with unique dependencies not out of the box.....
Building CI/CD code pipelines with the #AWS CDK. So much easier to synthesize the AWS CloudFormation templates here with the Lambdas in the same solution. #DevOps