$ devetra --intro
We migrate platforms to Azure, build infrastructure from nothing in Terraform, keep pipelines deploying without drama, and put AI agents on the tedious half of the work.
✓ azure · aws · terraform · kubernetes · gitlab ci/cd · mcp // riga, eu-wide
What we do, learned on government programmes, enterprise migrations and platforms serving hundreds of projects.
Agents and MCP connectors wired into daily delivery: writing and reviewing Terraform, diffing plans before they reach a pipeline, automating the steps nobody wants to repeat. Faster hands, same engineering judgement.
Planned, rehearsed and executed with both environments live until cutover is safe. Databases included: replication, failover, the full data load.
Reusable Terraform modules, landing zones and network topology your teams can build on without raising a ticket. Reproducible environments: one command up, one command down.
Pipelines that carry code to production on their own: build, test, deploy, roll back. One set of templates, not a copy per project.
Monitoring, alerting and incident response for platforms already in production. We learned what breaks, and how fast, supporting more than 200 concurrent projects.
Client names stay confidential. Results don't have to.
A multi-application migration with both clouds running in parallel until cutover. Reusable Terraform layer, one set of GitLab templates for two clouds, secrets moved to Vault. Application teams now provision their own environments instead of waiting on tickets.
End to end: hub-and-spoke topology, isolated private networking, three environments buildable and destroyable with a single command. PostgreSQL moved with cross-region replication and automated failover.
Inside the SRE team of a European IT services provider: Terraform standards other teams adopted, AKS monitoring managed through Helm, secure networking from firewalls to site-to-site tunnels, landing-zone compliance passed on the first review.
Architecture from a blank page: topology, environments, Data Factory, SQL Managed Instances, Front Door. AI agents and MCP connectors carried the repetitive work through the whole delivery, from sprint planning to monitoring.
Four phases. Each one ends with something you keep, whether we continue or not.
stage 1/4
We read your infrastructure, costs and risks, and put the findings in writing. A plan you could hand to anyone.
stage 2/4
Architecture, network topology and a security model that survives review. Agreed before anything is built.
stage 3/4
Infrastructure as Code, pipelines, environments. Everything reproducible, nothing done by hand twice.
stage 4/4
Monitoring, support and a clean handover to your team. Or we stay and keep it running.
The short version. Tools we ship with every week.
# stack.yaml
clouds: [azure, aws] iac: [terraform, bicep] delivery: [gitlab-ci, azure-devops, github-actions] runtime: [kubernetes, aks, docker] data: [postgresql, azure-sql, elasticsearch] operations: [datadog, azure-monitor, vault] ai: [mcp, claude-code, cursor, copilot]
Devetra reads as Dev, new era: development for a time when AI sits inside the delivery loop, with DevOps folded into the name. That is the work: cloud infrastructure, migrations, automation, and agents doing their share of it.
The practice is led by Boriss Serikovs, a lead system engineer working since 2020 on government programmes, enterprise cloud migrations and SRE at scale. Based in Riga, delivering remotely across the EU.
$ certifications --microsoft
Need the application side too? For full product delivery we team up with partner software studios.
Two sentences about what you are building, running or moving is enough. You get a practical next step, usually within a day.
# direct line
+371 28211078