JIŘÍ NOVOTNÝ

Red Hat

Track talk

2:30 PM - 3:10 PM I TRACK 2 I 8 November 2024

About Jiří

I’m Jiri, and I’ve been working as a software quality engineer at Red Hat for nearly three enriching years. 
My interests extend beyond just coding; I’m also deeply committed to education and community engagement.
The project that holds the most personal significance to me is Debezium, which I’ve been involved with and deeply passionate about since I joined the team.
Additionally, I’m actively engaged in the Skodjob project, where I contribute to developing various tooling for long-running test environments on the Openshift platform. 
Within this project, I maintain the Database Manipulation Tool (DMT), a vital tool for generating database load and facilitating idempotent manipulation of data, particularly for testing purposes.

Speech title

Fusing AI and Debezium performance testing: Generating test scenarios with ML models

Speech description

In the ever-evolving landscape of quality engineering, the fusion of artificial intelligence and Debezium presents a great opportunity. Our proposed talk delves into this innovative approach, showcasing how AI augments performance testing to unveil bottlenecks within Debezium, thus enhancing the quality of the product. At the heart of our endeavor lies the development of a machine-learning model trained to generate various testing scenarios. We managed to create an advanced testing framework that transcends conventional methodologies. Throughout our journey, we encountered many challenges that shaped our approach to machine learning and performance testing. The limitations of small datasets posed significant hurdles, often constraining our machine-learning models’ robustness and generalization capabilities. Moreover, deploying machine learning platforms proved non-trivial, requiring careful consideration of infrastructure and integration with existing systems to ensure seamless operation in our environments. Additionally, conducting performance testing presented complex problems, from identifying suitable metrics to capturing the intricacies of system behavior under varying load conditions. Through our experiences in developing machine learning models and conducting performance testing, we’ve gleaned invaluable lessons that extend beyond technical expertise. We’ve come to understand the importance of data quality, iterative experimentation, and recognizing the limitations inherent in our models. Deploying machine learning models into production environments has revealed the complexities of integration and maintenance, prompting us to navigate these challenges with diligence and adaptability. Performance testing challenged our problem-solving skills, adaptability, and commitment to continuous learning. Moreover, we’ve honed our effective communication and collaboration abilities, recognizing the significance of articulating our findings. Our presentation will expose our comprehensive testing stack and ML model, showcasing their development process and their tangible impact.

Key takeaways from our talk include the realization that leveraging AI in testing is more accessible than perceived. Furthermore, we’ll show how our approach shifts the focus from measuring performance to actively seeking bottlenecks, leading to a refined understanding of system performance, and ultimately enhancing product quality.

Join us as we explore the synergy between AI and Debezium, enhancing performance testing and contributing to advancements in software quality assurance.

Nov 7 - 8

Palais Wertheim, Vienna
For general inquiries contact info@testingunited.com

Follow our social media platforms

Take advantage of our GROUP and ISTQB discount to save money!​

Limited seats available.

Follow our social media platforms

© Testing United 2018. All Rights Reserved Krone Consulting s.r.o.