DDD Perth 2024

There was again DDD Perth 10th event at beautiful Optus Stadium in 2024. The event was the biggest ever as I remember. Below are the talks I joined. I skipped all the 20 minutes or shorter talks as these are just scratching the surface session.

Exploring the Latest in Multimodal Generative AI

Presented general idea of how machine learning evolved and what is capable of at the current stage. Little hard to fully follow if you don't have a machine learning experience like me. Impressed that Nvidia data scientist flow to Perth for this.  

Mastering Full Stack with Azure: Effortless Infrastructure and Well-Architected Framework

Azure Advisor - Been using Azure for a few year by now, but haven't paid attention to Azure advisor, which is the first thing to check if you use Azure. Azure advise provides recommendations on Cost, Security, Reliability, Operational Excellence, and Performance. Should try infrastructure as code approach. Azure Resource Manager (ARM) Templates, Azure Automation, and Azure Bicep. Useful resource: By pillar - PSRule for Azure What I should try

1. Stop/scale down resources during weekends using ARM
2. Autoscaling resources based on usages

 

The End of Microservices

In many cases, 'Microservices' connects each other and strongly tangled, which is worse than monolith. This talk reminded my of 2 cases of my company which needs some improvement in terms of application architecture. e.g.#1.  App A -> DB1 -> Linked DB2 & Linked DB3. Moving DB 3 can unknowingly break App A. e.g.#2. App A -> Queue -> API B -> API C -> DB1. Many steps to reach to DB1. Hard to set up new environment. Hard to trace where is an issue when exception is thrown.   GraphQL: I Don't Know What It Is And Now I'm Too Afraid To Ask Allow to generate query at API consumer side rather than API contract is given by the server side. GraphQL can be useful to fix the over-fetching problem. Just thought if over-fetching problem is not too bad, creating simple variant of API end point will resolve the over-fetching problem with low effort.  

Leveraging LLMs to build Intelligent Applications

AI Application using LLM, RAG, LangChain, etc. can be easy to start rather than thought. Start with registering OpenAI API. e.g. LangFlow was a booth in the conference room. The service provides a drag and drop UI to set up a RAG chat bot system, etc.   There wasn't shocking talks this year. Mostly already been heard but haven't personally tried. Clearly the best take away from the tech conference is not a knowledge but motivation for learning and chances to chat with other passionate developers from other side.