< session />
100 Million Lines of Code, Millions of $ Saved, 0 Wasted Builds; AI-Powered CI @ Atlassian Scale
Tue, 22 April
Managing a monorepo with 100 million lines of code presents immense challenges in infrastructure costs, scalability, and build efficiency. At Atlassian, our high-volume CI/CD pipeline executes hundreds of thousands of tests daily, leading to significant redundancy and inefficiencies in the development workflow.
By leveraging AI and machine learning, we have transformed our Continuous Integration (CI) pipeline to dramatically optimize build times and resource consumption. Through predictive test selection, powered by static code analysis and ML models, we skip 50-95% of redundant tests per change—without compromising quality. This innovation has resulted in thousands of engineering hours saved and annual infrastructure cost reductions of over $4 million.
Beyond test selection, AI-powered tooling has also reduced manual effort in code migrations by 50%, allowing teams to focus on innovation rather than repetitive maintenance tasks. This session will provide an inside look at how AI/ML advancements in developer tooling can streamline CI/CD processes, cut costs, and enhance engineering productivity at an unprecedented scale.
Key Takeaways
- AI-Powered CI/CD Optimization – How Atlassian scaled Continuous Integration across a massive monorepo with AI/ML-driven efficiency.
- Predictive Test Selection – Learn how machine learning models analyze code changes to skip 50-95% of unnecessary tests.
- Cost & Resource Optimization – Discover how these innovations led to $4M+ annual savings in infrastructure costs.
- AI-Driven Code Migrations – See how AI-assisted tools reduced manual migration efforts by 50%, accelerating large-scale refactors.
- Scaling Developer Productivity – Insights into how AI and automation improve developer experience, efficiency, and CI/CD scalability.
Target Audience
- DevOps & CI/CD Engineers – Professionals managing large-scale build systems, looking to enhance efficiency and automation.
- Engineering Leaders & CTOs – Decision-makers seeking cost-saving AI innovations in software engineering at scale.
- Software Developers & Architects – Developers working in large monorepos who want to optimize build and test cycles.
- AI/ML Engineers – Experts exploring real-world applications of AI-driven automation in developer tooling.
< speaker_info />
About the speaker
Hari Lingamagunta
Head of Engineering, Atlassian
Hari is the Head of Engineering for the Dev Infra org, which helps build world-class App experiences for Atlassian's customers across Jira, Confluence, and Trello leading teams across the US, Australia, Europe and India. With over 18 years of industry experience, Hari has previously worked at companies like Amazon, Concur, and SAP, among others, and has worked in various domains such as Retail, Travel and Expense Management, Enterprise software, and cloud platforms. Hari is passionate about leading high-performance engineering teams and leveraging AI to improve their Dev Loop while also focusing on Reliability, Scalability, and Performance.
