← Work

Research · Artificial Intelligence · DevOps

Self-Healing CI/CD Pipelines Using AI

Sole Researcher & Engineer · B.Eng. Final-Year Research·2025 – 2026Research project

My Bachelor's Engineering research project, and one of the strongest demonstrations of my software engineering and AI capabilities. It explores how AI can make CI/CD pipelines autonomous by detecting failures, identifying root causes, and automatically recommending or executing recovery actions, reducing developer intervention and increasing deployment reliability.

01Pipeline failure
02Collect build logs
03Analyze logs
04Classify root cause
05Select recovery strategy
06Execute recovery
07Validate & restore

Problem

Modern CI/CD pipelines regularly fail because of dependency conflicts, infrastructure failures, configuration mistakes, flaky tests, build failures, and deployment issues. These failures consume engineering time because developers have to manually inspect logs and work out the right fix.

Solution

Designed a self-healing pipeline framework capable of detecting failures automatically, collecting and analyzing build logs, identifying probable root causes, selecting an appropriate recovery strategy, validating that recovery worked, and returning the pipeline to a healthy state.

Tech stack

PythonDockerGitHub ActionsJenkinsKubernetesMachine LearningLog Analysis

Key features

  • Automated failure detection
  • Intelligent log analysis
  • Root cause classification
  • AI-assisted recovery
  • Rule-based remediation
  • Recovery validation
  • Modular architecture

Engineering highlights

Software ArchitectureArtificial IntelligenceAutomationDevOpsDistributed SystemsReliability Engineering

Impact

Forms the core of my B.Eng. research work: a framework for reducing manual log inspection and mean-time-to-recovery in CI/CD pipelines through AI-assisted automation.