The advancement of DevOps and DevSecOps has resulted in the creation of platform engineering teams, necessitating DevSecOps platforms that integrate their diverse toolsets. CloudBees has evolved to address these needs recently acquiring Launchable, which uses AI to streamline software testing and accelerate development.

In this show, Shawn Ahmed, CPO at CloudBees, talks about the evolution of DevOps and CloudBees’ own journey as a company. Ahmed discusses the importance of having platform engineering tools that reduce developer’s time spent on non-coding tasks, how AI/ML is being used, and their acquisition of Launchable. Ahmed points out that numerous studies have consistently found that, on average, developers spend only around 20% to 35% of their time actually coding.

Evolution of DevOps and how CloudBees has grown as a platform

• CloudBees started as a Jenkins enterprise company but now offers a broad DevOps and DevSecOps platform, supporting various application types and deployment environments.
• Ahmed describes the evolution of DevOps from its origins in Continuous Integration and Continuous Delivery (CI/CD) to a broader platform that now includes extensive DevOps and DevSecOps capabilities.
• Ahmed discusses how software development has evolved from basic CI workflows to full lifecycle management with enhanced security, progressive delivery, and rapid feedback for faster and more adaptable releases.

The need for effective DevOps and DevSecOps platforms to reduce wasted developer time

• The evolution of DevOps and DevSecOps platforms centers on placing developers at the core, allowing them to maximize their coding time by integrating platform engineering teams that manage the entire software lifecycle.
• Platform engineering teams, comprising diverse roles, need collaborative platforms that integrate policies and processes to enable developers to move efficiently from coding to release in a week.
• Ahmed explains that platform engineering teams seek DevSecOps platforms that are highly adaptable, allowing integration of diverse tools and customization to meet their unique requirements, driving the evolution of these platforms to better support such collaborative needs.
• Developers spend only 20-35% of their time coding, with the rest taken up by managing issues like debugging and test triaging. Ahmed emphasizes using AI and machine learning to optimize testing processes to free up more time for developers to code.

How AI and machine learning are being used to free up developers’ time

• AI/ML can help to reduce developers’ wasted time by optimizing test processes. For example, predictive test selection can identify which tests are likely to fail, allowing developers to skip unnecessary tests and save time.
• Ahmed highlights innovations from companies like Launchable, which reduce test run times by up to 80%, exemplifying how AI can give developers more time to focus on coding.
• Large language models (LLMs) like OpenAI’s Copilot have made significant strides in code generation, proving highly valuable for developers. However, Ahmed believes their impact beyond coding, such as in areas like QA and testing, remains less clear.
• While code generation is well-established, applying AI to other stages requires careful consideration of the most beneficial use cases and ongoing evaluation of their effectiveness.

The value of AI-generated code and the need to identify use cases

• Ahmed emphasizes the value of AI in generating code and improving testing processes within organizations and among customers. While AI and machine learning show promise, their effectiveness in other software development aspects remains uncertain.
• Ahmed emphasizes the importance of identifying specific use cases within software development, such as QA and testing, to apply AI and ML effectively. Focusing on these targeted areas can address key challenges and provide substantial value.
• There is a need to identify use cases for applying AI in CI/CD pipelines, specifically in automating testing processes, handling code influx, and resolving pipeline failures.
• Customers are looking for solutions to manage the increased volume of code in existing pipelines. Ahmed discusses how AI and ML can assist in automating responses to pipeline failures, triaging issues, and supporting progressive releases.

How AI/ML is being integrated into developer cycles, and Cloudbees’ acquisition of Launchable

• As CI/CD tools like Jenkins continue to handle increasing volumes of code, companies are exploring how AI and machine learning can enhance their pipelines. Key areas for AI applications include improving QA, triaging pipeline failures, and optimizing progressive releases.
• By leveraging AI to analyze failure patterns, automate issue resolution, and manage feature rollouts incrementally, companies can streamline processes and reduce manual intervention.
• Ahmed explains how integrating AI/ML into development cycles streamlines processes by enhancing progressive delivery and improving the efficiency of triaging pipeline failures, thus optimizing response times and overall workflow efficiency.
• Ahmed talks about how the acquisition of Launchable exemplifies this approach, offering advanced capabilities that integrate with existing pipelines to enhance efficiency and accelerate innovation.

Guest: Shawn Ahmed (LinkedIn)
Company: CloudBees (Twitter)
Show: Let’s Talk

This summary was written by Emily Nicholls.