STAREAST 2023 - Industry Technical Presentations

COVID-19 has accelerated the development of contactless solutions. With bots/conversational interfaces being the most preferred way of quickly deploying self-service tech, the popularity of ChatGPT has proven that conversational interfaces are the most natural way of interacting with apps and the internet. Key aspects of any AI-based bots are their ability to understand the user’s intent and provide the right solution.

In a changing environment, successful software is continuously evolved to meet the users‘ needs. However, changes to the software increase the complexity and make subsequent changes ever-more difficult and – in the long run – the software impossible to be maintained or improved any further. Stopping this "software erosion” is crucial for sustainable quality and a long life-cycle of the software. Erosion-free software ensures that business critical safety and security standards are met.

The drive to transform pushes application delivery teams to release faster. Yet enterprises need to deliver quality products and services to woo customers and gain market share. Consequently, quality is often lost or even sacrificed as efforts to increase delivery velocity are prioritized, and the alignment between business and development is lost. But software quality should maximize the value of your enterprise strategy—not put it at risk. And it isn’t a question of speed vs quality; rather it's how we can do both at the same time. 

Learning Objectives: 

To keep up with fast paced delivery without sacrificing quality or having armies of testers, effective tooling is a necessity. Off-the-shelf tooling often doesn’t provide the impact required and custom tooling from scratch requires large investments. Targeted, built-for-purpose tooling can balance both worlds - maximizing impact and finding innovative solutions to *your* problems. Quality Engineering means taking testing to the next level and looking beyond how things have always been done - including how tools can transform the way you test.

Learning Objectives:

Many quality teams measure success based on the number of scripts written and deployed, that is Test Coverage. These have been practical metrics given the time constraints for manual testing and for creating and maintaining test automation. But now that AI is providing the opportunity for teams to achieve massively greater efficiency through autonomous testing, these measures fall short. Application Coverage is a more complete measure of the effectiveness of your testing program yielding more issues more quickly.

Learning Objectives:

A typical Kubernetes application is usually made of several containers, and services within those containers. These are usually sat behind an Egress/Ingress controller and many security layers so that interactions with the application are restricted to the overall application APIs, whether they are published or not. Access to the internal services of the applications are largely restricted though testers might need to access internal services, and this is all what the transparent remoting feature is about.

Today, everybody is tasked with integrations. We must integrate our product with EVERY system, component, and operating system under the sun. Unfortunately, it might not be desirable or even possible to access those systems during development. There could be security, performance or maintenance issues that make them unavailable. We see teams having dependencies on services owned by other groups which are not accessible, which halts development.

This is when we mock, stub, or virtualize.

All too often, organizations buy into the idea that the key to long-term, successful automation is a straightforward combination of manpower and an automation tool. But it’s not as simple as that. In order for automation to be truly scalable in the long term, teams must establish strong collaboration and work together to define their overarching strategy and approach.

Artificial Intelligence (AI) is revolutionizing how software testing is performed, allowing for faster and more accurate detection of defects. However, with this new technology comes new risks that must be managed to ensure the quality and reliability of the software being developed. In this talk, Adam Sandman will explore the concept of Risk-Based Testing (RBT) and how it can be applied in the age of AI. He will discuss the challenges that arise when implementing RBT, including the need to balance the benefits of AI with the risks of false positives and false negatives.

Autonomous Testing is rapidly becoming a trendy new word in test automation, but what does the term mean? The definition has important implications in ascertaining whether we are, or it is possible, to be close to autonomous testing or not. The most basic definition could be “an autonomous system needs to be able to make decisions for itself.” This requires data. Lots of data. This brings us to an interesting question: from where do we get this data? Production, and that has its own problems.

With the emergence of deep learning integrated into today's software products, a modern software development process needs an effective way to automatically test and deploy machine learning models as part of product delivery. In this presentation, Coveros CEO Jeffery Payne introduces deep learning AI models and how to effectively test and deploy them as part of product development. Approaches to testing models will be discussed as well as how to automate your deployment process when machine learning is part of your product.

Join this session to: