Dinesh Sai ← All projects

Case study · DeliveryProduction, MedTech

Release pipelines that prove every build is safe to ship.

CI/CD for medical and scientific software, where a release has to pass quality, security and governance checks, not just compile. Jenkins and GitHub Actions pipelines, Helm-based delivery and release automation across dev, staging and production.

  • Role: Lead DevOps Engineer
  • Jenkins · GitHub Actions · Argo CD
  • Sole DevOps engineer for SCIEX at IDC
kubernetes-java/Jenkinsfile (public lab)
stages {
  stage('Checkout')       stage('Lint')
  stage('Build')          stage('Test')
  stage('Code Coverage')  stage('SonarQube')
  stage('Build Docker Image')
  stage('Trivy Scan')     stage('Push Docker Image')
  stage('Deploy to Kubernetes')
}
post { always { junit …; publishHTML … } }
Stage names from a public lab pipeline: quality and security checks before anything ships.
80%faster time-to-deploy
50%less manual release effort
25%better availability
40%less frontend release friction

Results from my work at Danaher and Ajmera Infotech, as listed on my resume.

The problem

In regulated software, "it built" is not a release.

A medical-device release needs evidence: which code went out, what tested it, which scans it passed, and who approved it. When those steps are manual, releases are slow, inconsistent between environments, and painful to audit.

The goal was pipelines that produce the evidence as a side effect of shipping, with automated checks in a fixed order, notifications and approvals in the tools teams already use, and the same path from dev to production.

What I built

From commit to production.

  1. Jenkins pipelines wired into the workflow

    At HemoCue, Jenkins integrates with Bitbucket and Jira to automate build, test and release, which improved uptime and availability by 25%. At SCIEX, Slack and Jira integration adds real-time release notifications, automated approvals and end-to-end traceability.

  2. Security in every release

    SCIEX pipelines run Veracode SAST and DAST plus Cybellum binary analysis, so security and compliance checks gate every release. The DevSecOps case study goes deeper.

  3. Helm delivery and a one-run installer

    A Python installer provisions AWS infrastructure and deploys the application with Helm in one run, which cut time-to-deploy by 80%. Argo CD with Kustomize then gives auditable, environment-specific rollouts with one-click rollback.

  4. Release automation

    Python and Bash automate release management and environment provisioning, with identical pipelines across dev, staging and production. Manual release effort dropped by 50%.

  5. Quality gates on GitHub Actions

    At Ajmera Infotech I built CI/CD on GitHub Actions with SonarQube quality gates and ECR image lifecycle management, AWS Amplify for frontend deployment and AWS Device Farm for mobile QA. Frontend release friction fell by 40%.

Before

Earlier delivery work.

Subscribe IT · fintech

Automated Liquibase database schema migrations in the pipeline, synced GitHub Actions with Jira, and integrated Opsgenie alerting, which cut incident resolution time by 35%.

CSS Corp · ad tech

Built Jenkins pipelines that allowed several zero-downtime deployments an hour, and ran Docker and Docker Swarm clusters, improving service resilience by 30%.

In code

Pipelines you can read.

Stack

Tools used

Want the details?

I can talk through pipeline design, release governance and the installer in more depth, within what my employer allows.

Get in touch More projects