Function Hall

List of Events

Here is a list of event, includes speaker session and other intermediate slots (e.g. lunch):

Start using the Kubernetes Gateway API now rather than Ingress

Have you been struggling to set up network load balancing and managing web traffic in Kubernetes?

The Ingress API in Kubernetes seems to be too basic and not expressive enough to help you accomplish what you need. You may end up with many downstream implementation details in your configuration that are not portable. It gets even more complicated when you have developers, operators and network engineers working on the same shared configuration.

Dynamic Kubernetes based Feature Environment on AWS

The core engineering principle at Prenetics is “You build it; you run it” and in order for our engineers to fully embrace this DevOps model, they must be able to easily build and test locally, thereby reducing the build and test feedback loop to the minimum. To this end we have developed the Dynamic Feature Environment allows engineers to spin up/down a fully functional independent K8s cluster that runs our full set of microservices. In this talk we will share our experience implementing and using the Dynamic Feature Environment.

結束一台獨大：去中心化社群媒體的內容審核難題

Facebook、Instagram、Twitter等社群媒體一台獨大。有人詬病它們監督不力，讓虛假資訊和極端思想滿天飛，亦有人不滿它們審查言論和無故封鎖帳號。有人因此轉投MeWe等另類中心化平台，有些人則希望結束一台專政，建設去中心化平台。

但是去中心化只是手段，更好的社群才是目的。去中心化平台該如何處理不良內容，克服內容審核的不可能三角？去中心化社群媒體，很多人都聽過Mastodon，那你有聽過近期流行的Bluesky Social 嗎？Mastodon有什麼不好要去開發新的網絡協定？社群媒體演算法對用戶是壞東西？

從Mastodon到Bluesky Social，講者會分享去中心化社群媒體面對的內容審核難題，從技術角度深入解構兩個系統架構上的取捨。可是到最後，會發現最難的從來不是技術問題，還是人的問題。


 

How to build a powerful AI VR Virtual Assistant with AWS AL/ML service and OpenAI?

AWS AI Assistant is a serverless applications and integrates most of the AWS AI/ML services and Open AI.
The most popular tech news in 2023 is the raise of generative AI with Large language models. With those powerful pre-built model or AI API, we can build a very powerful virtual assistant with the combination and integration of all AWS AI/ML services with OpenAI API or Hugging Face pre-trained model. For my use case, I hope my students learn and use AI but I do want to have certain control and monitoring. So, we want to make it fun and powerful for my students.