Rebuilding Jira CI Reliability
Reworked Jira CI through a hybrid approach to reliability, cost, and pipeline speed.
- CI Reliability
- Developer Experience
- AI Automation
- Platform Engineering
Nitin Goel
Senior Software Engineer at Atlassian





From their time as

Senior Software Engineer
Atlassian • 2025
Overview
Nitin rebuilt Jira CI reliability while reducing both operating cost and pipeline runtime. He used a hybrid deterministic and AI-based classifier to improve build-failure handling.
The Story
Nitin worked on Jira CI at a point where reliability, cost, and pipeline duration were central developer experience concerns. The mandate was to improve the platform without introducing costly or slow automation.
He designed a hybrid build-failure classifier. Deterministic static analysis handled clear feature-flag and syntax cases, while an AI classifier evaluated ambiguous code paths and contextual flag use.
The trade-off was deliberate. A fully AI-based path would add API cost and CI latency, so Nitin expanded deterministic coverage to keep AI use focused on ambiguity.
The resulting platform work rebuilt Jira CI reliability from 56% to 99.6%, cut CI cost by 44%, and reduced runtime from about 80 minutes to under 30 minutes.
Ownership Snapshot
Broad role
Senior software engineer on Jira platform work.
Goal
Improve CI reliability and developer workflow efficiency.
Direct ownership
Designed the hybrid classification approach and contributed to implementation.
Team execution
Worked with Jira engineering teams on platform adoption and operations.
