Deploying since 2007

Borys Babii

Senior DevOps & Cloud Engineer

Borys Babii is a Senior DevOps & Platform Engineer at DataArt, based in Cherkasy, Ukraine, with 15+ years in IT and 10+ years in DevOps and cloud engineering. He works across an enterprise AWS estate of 60+ accounts — splitting, migrating, automating and cost-optimising it — and makes shipping to it boring.

15+ years in IT · 10+ in DevOps · Cherkasy, Ukraine

01 The pipeline

Commit to production, with no one holding their breath. Scroll to run it — and scroll back to rewind. Every stage is something I have built, run and been paged for.

#482 main Running 0s
  1. Checkout git fetch 4s
  2. Unit tests pytest, go test 31s
  3. Linters ruff, golangci-lint, tflint 12s
  4. Static analysis SonarQube, tfsec 26s
  5. Integration tests docker compose 1m 48s
  6. Deployment Blue/Green, canary 2m 06s
  7. E2E tests Playwright 3m 21s
  8. Load tests k6 4m 10s

$ git fetch --depth=1 origin main

02 Zero-downtime by default

Blue/Green means the new version is already running and healthy before a single user reaches it. If the metrics turn, traffic goes back — automatically, in seconds, with no rebuild.

  • Two identical environments, one live at a time
  • Health checks gate the switch
  • Automated rollback on error-rate or latency
  • Same path for Python, Go and Node.js services

03 The toolbox

Cloud

AWS — EC2, ECS, Lambda, RDS, VPC, CloudFront, Transit Gateway, WAF, ALB, S3. GCP.

Infrastructure as Code

Terraform — modules, monorepos, Atlantis, Atmos. CloudFormation, Packer.

Automation

Ansible, Chef, Jenkins shared libraries in Groovy.

CI/CD

Blue/Green, GitOps, Jenkins, Octopus Deploy, GitHub Actions, Travis, Circle CI.

Observability

Datadog, Grafana, New Relic, Splunk, Graylog, Sentry, Nagios, Zabbix, PagerDuty.

Languages & containers

Python, Bash, Go, PHP, JavaScript, TypeScript. Docker, Compose, Swarm, Vagrant.

Platform engineering

AWS Organizations, multi-account landing zones, account bootstrap automation, internal CLI tooling, cost attribution.

AI tooling

Claude Code skills and agents for recurring platform work — reviews, migrations, runbooks.

04 Mission log

Scale

A 60+ account AWS estate

Splitting legacy accounts that had accumulated many services into smaller accounts, one per team or per service. Clear blast radius, clear ownership, clear bills.

AWS Organizations · Terraform · Atmos

Platform

Account bootstrap, automated

A new AWS account arrives with its baseline already in place — guardrails, roles, networking, logging, monitoring. No runbook, no week of clicking.

Terraform · AWS Organizations · CI

Tooling

Internal CLI for module versions

Built the CLI that tracks and bumps internal Terraform module versions across repos, so upgrades stop being an archaeology exercise.

Go · Python · Terraform

Cost

Continuous cost optimisation

Ongoing right-sizing, waste removal and spend attribution across the estate — treated as routine engineering, not an annual panic.

AWS Cost Explorer · Terraform · Datadog

Migration

Single account to multi-account

Moved a whole production stack from one AWS account into isolated accounts. Zero downtime, no maintenance window.

Terraform · AWS · Jenkins · Ansible · Octopus

Immutable

Golden image pipeline

Reusable AMI workflows, so every release ships a fresh, tested image instead of a patched server.

Packer · Ansible · Jenkins

Delivery

One path for three runtimes

Shared Jenkins libraries that deploy Python, Go and Node.js services through the same reviewed path.

Jenkins · Groovy · Docker · Make

Observability

One pane of glass

Unified dashboards, alert routing and log aggregation across a fleet that previously had none.

Datadog · Grafana · Splunk · Sentry · PagerDuty

05 Trajectory

  1. 2019 – present

    Senior DevOps & Platform Engineer — DataArt

    Platform engineering across an enterprise AWS estate of 60+ accounts. Splitting legacy accounts that carried many services into smaller accounts per team and per service, automating account bootstrap, and running a continuous programme of migrations and cost optimisation.

    Built internal CLI tooling for Terraform module version management, Blue/Green deployments with automated rollback, and end-to-end monitoring and log aggregation. Currently developing Claude Code skills to automate recurring platform work.

  2. 2013 – 2019

    Senior DevOps Engineer — MEV, LLC

    Led the migration from IBM SoftLayer to AWS, designing the scalable and highly-available architecture behind it. Built Groovy pipelines and shared libraries, Docker Swarm clusters, a Docker registry and custom golden AMIs. Ran Chef server, Nexus, Splunk and SonarQube, and covered the infrastructure with tests.

    Introduced an immutable infrastructure approach with Packer and Ansible, and kept developer, QA and customer environments supplied throughout.

  3. 2011 – 2013

    Web Software Developer — MEV, LLC

    Built PHP healthcare and e-commerce solutions with Zend Framework and WordPress.

  4. 2007 – 2011

    System Administrator — Megastyle ISP

    L1/L2 support, network and server configuration.

06 Shipped & written

Code

Writing

28 July 2026 · Medium Docker Layers: One Dockerfile, Traced All the Way Down Which instructions really create layers, how copy-on-write works, and how to check every claim yourself.

Education

  • ISMA, Riga — Professional Master's, Computer Systems (since Sept 2024)
  • East European University of Economics and Management, Cherkasy — BSc, Economic Cybernetics (2016–2018)
  • Polytechnic College, Cherkasy — Software and robotic systems (2003–2005)