The annual AWS Summit recently rolled into Sydney, bringing together thousands of cloud enthusiasts for (up to) 3 days of learning, networking, and looking at the latest innovations from AWS, partners and customers.
Tuesday kicked things off with the partner day, allowing AWS partners to connect and bring everyone up to date on various partner programs available to assist customers building on AWS. As an AWS Ambassador, I attended the annual AWS ANZ Ambassador appreciation night at Braza restaurant where we connected and caught up with each other. We have representatives from most parts of Australia so it’s one of the only times we can catchup in person each year.
Day 1: Builders Day
The big news on the opening day keynote was the launch in Australia of Amazon Bedrock, AWS’ foundation model service. While not all functionality is available at launch, Australian organisations and developers now can start experimenting with generative AI models, especially for locally regulated industries who are unable to take advantage of international AWS regions.
As a reminder, while Amazon Bedrock is a service that provides managed access to Foundation models, with a bit of effort you can also leverage SageMaker as an alternative so it’s always worthwhile to compare the ongoing costs. Bedrock is great for quickly trying things out, but as with all pre-production planning be sure to model your costs prior to production deployment carefully.
The day featured sessions from companies showcasing their use of generative AI. Leonardo.ai as an organisation grew from 6 to 100 employees in 12 months and have customers in over 180 countries! They shared how they use AI to produce 6 million digital assets per day for gaming, retail, marketing, real estate and other sectors. Atlassian demoed their Atlassian Intelligence offering which allows natural language search across development tools like Jira, Confluence and BitBucket plus third party applications like Figma and GitHub.
Day 2: Innovation Day
A highlight of the Day 2 keynote for me was a fascinating history lesson by Professor Genevieve Bell, the Vice Chancellor of ANU. She spoke about the construction in the 1870s of the Overland Telegraph Line from Port Augusta to Darwin, which allowed Australia to finally communicate in real-time with the rest of the world instead of waiting about 45 days for news by ship.
This history lesson tied into Amazon’s ambitious Project Kuiper initiative to launch 3,236 satellites to provide affordable broadband internet access globally, with Australia set to be one of the first demo locations in 2025.
Other highlights included Swimming Australia explaining how they use AWS machine learning and computer vision to analyse training and optimise for peak performance. Their dashboards provide coaches rich data to fine tune training programs, and with this ML assistance helping propel Australia’s swimmers to success at meets like the upcoming Paris Olympics.
Personal Session Highlights
- Build customised business solutions using foundation models is awesome in demoing SageMaker JumpStart and how it can be used with the various Foundation models.
- Building smart applications using large language models is a great session to watch to give an overview of building with Bedrock. Frank also goes over the different use cases when you need to use prompt engineering, augmenting LLM’s with Retrieval Augmented Generation (RAG), augmenting the LLM by using Agents and finally, when to use fine tuning and continuous pretraining.
- Mastering serverless: A day in the life of a developer is a masterclass on how to build cloud-native serverless solutions.
- AWS Super Session: Build without limits: The next-generation developer experience at AWS was a supersession for developers where among other things AWS announced a Project Development Kit (pdk) to assist developers getting started on new projects with some standard boilerplate frontend and backend code and deployment pipelines.
- Project Kuiper in depth session goes through in more detail the problems solved by the technology and use cases.
Summary
It was great to see the huge attendance again after a quiet few years. Yes, there was a predominant focus on GenAI but there were also sessions on “traditional” topics around modernisation and data.
If you takeaway only one thing from Sydney Summit this year, let it be this: Know your data. Know where your data is and make it available to your organisation. Understand your business problem. Only then can you take advantage of not only the newest GenAI innovations, but also the many traditional Machine Learning offerings that have been available for years – it ultimately depends on your business problem what tools you use to solve it.
If you need help unravelling your data or making sense of the latest AWS technology, please contact Cevo – we are here to assist.
