NameRoman Chuprikov

RoleHead of DevOps / Platform Engineering

StatusRemote

Contactroman@chup.cc Blogblog.chup.cc

I build platforms where teams ship faster — and sleep better.

I lead Platform Engineering, develop services in Python, and use AI in my engineering work.

ChangeDeliveryProduction

Discuss a challenge See outcomes

1M+ / monthplatform audienceweekly → several per hourrelease frequencyhours → minutestime to market

Changes you can see in the metrics

I do not count a deployed tool as an outcome. The outcome is when teams ship changes faster while the system stays manageable.

Release frequency

Deployment frequency

Before: once a week
After: several per hour

Product teams ship without waiting in a week-long release queue.

Change speed

Time to market

Before: hours
After: minutes

Changes reach production faster and more predictably.

Operational control

IaC coverage

Verified coverage: 100% of server infrastructure

All server infrastructure is defined as code and reproducible.

1M+ / monthplatform audiencehundredsservers across several data centersseveralKubernetes clustershundredsservices in production

How I build platforms

I start with the bottleneck in the flow of change, not with Kubernetes. Then I standardize the path, automate it, and close the feedback loop.

Standardize

Golden paths, shared templates, and self-service replace a custom process for every service.

Control metricsInfrastructure in IaC · adoption of standard pipelines

Automate

Pipelines, GitOps, and infrastructure as code shorten the path to production.

Control metricsRelease frequency · time to production

Observe

SLOs, metrics, logs, and alerts show user impact, not just server state.

Control metricsServices with SLOs · actionable alerts

Improve the system

Incidents end with a postmortem, action owners, and platform changes.

Control metricsCompleted incident actions · recurring incidents

How I lead teams

I led a 12-engineer DevOps/SRE team. I provide context and autonomy while keeping responsibility for outcomes, architecture, and incidents explicit.

Connect work to the product

I translate technical initiatives into delivery speed, reliability, and the cost of change.

Control metricsRelease frequency · time to production · unplanned work

Establish ownership

Every service, decision, and postmortem action has an owner and a clear definition of done.

Control metricsServices with owners · completed commitments

Make risks visible

The team sees constraints, decision costs, and speed-versus-reliability trade-offs early.

Control metricsOpen risks · overdue actions

Multiply autonomy

I share context, grow leaders, and reduce the system’s dependence on individual people.

Control metricsWork completed without escalation · ownership coverage

Three areas of responsibility

Twenty years in infrastructure, from systems administration to leading a DevOps function. In every area, the point is not the technology itself but a better way of working.

Travel tech

Head of DevOps

Scope: The DevOps function and delivery processes for product teams.
What I changed: Created a shared CI/CD approach and made delivery ownership explicit.
Outcome: A shared CI/CD standard for product teams.

Services marketplace

Delivery transformation

Scope: The speed and predictability of the path from change to production.
What I changed: Rebuilt pipelines and the release process around automation and feedback.
Outcome: Releases moved from once a week to several per hour. Time to market fell from hours to minutes.

High-load e-commerce

DevOps / Platform

Scope: The platform for a high-load product and infrastructure across several data centers.
What I changed: Evolved the Kubernetes platform and moved server infrastructure management into code.
Outcome: 1M+ visitors per month, hundreds of services, and 100% of server infrastructure in IaC.

Code and tools

I choose the stack to match the constraint and the metric. A tool cannot replace architecture, process, or an owner accountable for the outcome.

AI in engineering

I use AI as a working tool in everyday engineering tasks.

Services in Python

I develop Python services for automation and platform tasks.

Platform

Kubernetes · Helm · Terraform · Terragrunt · Ansible

Reproducible infrastructure and self-service for product teams.

Control metricsInfrastructure in IaC · adoption of standard templates

Software delivery

Argo CD · GitLab CI · Jenkins

A short, controlled path from commit to production.

Control metricsRelease frequency · time to production

Observability

Prometheus · VictoriaMetrics · ELK · Grafana

Signals that show user impact and help teams find root causes quickly.

Control metricsServices with SLOs · actionable alerts · recurring incidents

Cloud and data centers

AWS · Yandex Cloud · on-premises

Architecture that does not depend on a single hosting model.

Control metricsCapacity headroom · infrastructure in IaC · cost