Freelance Work | Mirac Can Yılmaz

I take freelance work on Kubernetes and CI/CD

I have been a DevOps engineer since 2022. I have built Kubernetes clusters, written CI/CD pipelines and run GitOps migrations for teams in finance, betting and aviation. Most recently I built a 3-master, 14-worker Talos cluster on Hetzner with Terraform and run it with ArgoCD, Vault and Keycloak.
I take one project at a time, so each one gets my full focus. Tell me what you need and I will say plainly whether I can take it and how long it will take.

Timeline Set after we analyse the project

Response time 24 hours

Time zone UTC+3 (Türkiye)

Availability Immediately

Invoicing Business invoice

Working style Remote

Problems I've solved

Problems I ran into in production, what I did, and the measured result. If you're facing something similar, I can do the same work in your environment.

All code is delivered to your environment: Terraform, Helm and pipelines stay in your repository, along with runbooks.

Enterprise software · 2026

CI stayed in Azure DevOps, CD moved to Argo CD

Problem: Builds ran in Azure DevOps, but deployments reached each environment a different way; it was unclear which version was running where.
What I did: I installed Argo CD on Rancher-managed clusters. The Azure DevOps pipeline builds the image and writes the tag to Git; Argo CD syncs the cluster from Git.
Result: Releases became traceable and consistent across environments.

Argo CDAzure DevOpsRancherKubernetes

Set up Azure DevOps + Argo CD for you Read the lab post

Online betting platform · 2025–2026

Company-wide GitOps on a highly available Argo CD

Problem: Deployments were pushed from pipelines into the cluster; Argo CD depended on a single node and teams had no shared layout.
What I did: I installed Argo CD in high availability, made the architecture decisions for company-wide GitOps, documented the setup and trained the teams.
Result: Teams run their own deployments through one GitOps layout.

Argo CDGitLab CIKubernetesHelm

Let's build this in your environment Read the lab post

Freelance · 2026

Secrets moved out of Git and into Vault

Problem: Application passwords lived in Git and in manifests; changing one was a manual, risky job.
What I did: I set up HashiCorp Vault and External Secrets Operator; applications kept reading the same Kubernetes Secret name while the source became Vault. Cluster login moved to Keycloak OIDC.
Result: No secrets left in Git; rotation happens in one place.

HashiCorp VaultExternal Secrets OperatorArgo CDKeycloak

Move your secrets to Vault without downtime Read the lab post

Freelance · 2026

Production Kubernetes from scratch: 17 nodes

Problem: The team had no Kubernetes platform managed as code that could be rebuilt on demand.
What I did: I provisioned a Talos cluster with 3 control plane nodes and 14 workers on Hetzner Cloud with Terraform, load balancer included, and handed it over with Argo CD, Helm, Grafana and Loki.
Result: The whole cluster can be rebuilt from one Terraform and Git repository.

TerraformTalosHetzner CloudArgo CDGrafanaLoki

Let's build this in your environment Read the lab post

Online betting platform · 2025–2026

Container scanning: from zero to 100+ services

Problem: No container image in the company was scanned for vulnerabilities.
What I did: I designed a V2 scanning architecture with Trivy, added it to the CI pipelines and reported critical findings to senior management.
Result: Image scanning went from no coverage to 100+ services.

TrivyGitLab CIKubernetes

Let's build this in your environment

Banking · 2024–2025

150+ repositories from TFS to GitHub, by script

Problem: Hundreds of repositories sat in legacy TFS; moving them by hand would take months and leave repositories without owners.
What I did: I wrote a Python migration tool, connected SonarQube through its API and generated a CODEOWNERS file for every repository.
Result: 150+ repositories moved, each with a review owner from day one.

PythonGitHubTFSSonarQube

Let's build this in your environment

Engagement Models

I work under different models depending on the shape of the job; we decide together which one fits during the intro call.

Each package shows a starting price; I send the final price in writing by email once the scope is clear. For work that needs continuity beyond the packages, I also offer monthly fractional support.

Project-Based Build

A defined scope and a clear delivery schedule.

Kubernetes cluster build from scratch (Talos, kind, Rancher)
Cloud infrastructure and load balancer provisioning with Terraform
GitOps rollout with ArgoCD and hand-written Helm charts
Runbook and architecture documentation at handover

Best for: Teams building or rebuilding their platform from scratch.

Pricing: Fixed-price quote by email once the scope is clear.

Infrastructure Review

A review of your current setup, turned into a prioritised roadmap.

Review of clusters, pipelines and deployment workflows
Findings report covering security, cost and resilience
Improvement backlog ranked by impact and effort
Closing session to walk the team through the findings

Best for: Teams already running who want to know what to fix first.

Pricing: Fixed one to two week scope; fee shared by email.

Monthly Retainer (Part-time DevOps)

Reserved weekly hours for teams without their own DevOps engineer.

A fixed number of reserved hours each week
Pipeline maintenance, release workflows and infrastructure changes
Observability and alerting improvements
Day-to-day support for the team's questions

Best for: Teams needing ongoing support without a full-time hire.

Pricing: Monthly fee based on the weekly hour commitment, agreed by email.

Training & Team Enablement

Hands-on enablement so your team can own what was built.

Hands-on sessions on Docker, Kubernetes, CI/CD and GitOps
Exercises built on the team's own codebase
Onboarding docs and runbooks
Optional one-to-one mentoring

Best for: Teams that want to run the handed-over system themselves.

Pricing: Priced per session; fee shared by email.

Areas I Can Help With

Things I have built and operated in production myself.

Kubernetes Platform Builds

Production-ready cluster builds, upgrades and capacity planning.

TalosRancherHelmKustomizekind

GitOps & Release Management

Deployments driven from Git: traceable and easy to roll back.

ArgoCDHelmfileKustomizeGit

CI/CD Pipeline Setup

End-to-end build, test and deploy pipelines, self-hosted runners included.

GitHub ActionsGitLab CIJenkinsAzure DevOps

Infrastructure as Code

Version-controlled, reproducible infrastructure instead of hand-built servers.

TerraformAnsiblePackerHetzner CloudGCPAWS

Observability

Metrics, logs and alerts in one place, without the noise.

PrometheusGrafanaLokiELK StackOpenTelemetry

DevSecOps & Secret Management

Getting secrets out of repos, image scanning and centralised authentication.

HashiCorp VaultExternal Secrets OperatorTrivyKeycloak OIDCLDAP

Developer Experience

Project templates, internal portals and repo migrations to speed teams up.

BackstageCookiecutterNexusSonarQubePythonGo

How We Work Together

Intro call

A free 30-minute call to go over your current setup, your team and what you expect.

Scope and quote

I put the work, the deliverables and the schedule in writing; once we agree, we start.

Implementation

Everything lands as code (Terraform, Helm, Ansible) in your own repository, with a short weekly status note.

Handover and documentation

I hand the system over to your team: runbooks, architecture notes and a walkthrough session if needed.

Payment and invoicing

Scope and price are fixed in a written proposal; payment and invoicing details come with it.

Invoiced work

Every payment comes with a business invoice.

Payment methods

Bank transfer (Türkiye) TRY · QNB
International transfer (SWIFT) USD, EUR · QNB
Card payment Buy Me a Coffee · USD

Proposal and delivery

Scope, deliverables and timeline are in the written proposal. Bank details are shared with it; all code and documents (PDF) are delivered to you.