Yevhenii L.

Senior DevOps & SRE for production systems: AWS/GCP, Kubernetes, Terraform, Linux. I stop DDoS attacks, fix outages and runaway cloud bills, and run self-hosted AI on your own GPUs. Top Rated, 63 Upwork jobs, 5,300+ hours, 100% Job Success. Hi, I'm Yevhenii. I've been in IT since 2017. Most of my work starts the same way. Something is broken or costs too much, and nobody knows why. A store was getting attacked on checkout and payments still had to work. I set up Cloudflare rules and rate limits, and the biggest attack ended with 0 server errors. A client's AWS bill almost tripled in one month. I found the exact hour it started. Alerts in a camera system came about 10 seconds late. It looked like a frontend problem. It was the order of steps in the pipeline. A monitoring stack looked fine but nobody was getting alerts. Now it checks servers, leads and sales every minute. If sales stop, you hear about it even when every server is green. An app moved off Replit looked finished on the new server. Then I tested it in a real browser and login was blocked with 403. The fix was small, but nobody could have logged in. A client got a scary report from a security scanner. I checked each finding by hand. Some were real, some weren't, and they knew what to fix first. What people usually hire me for: Server or site is down and you need it back today. Cloudflare setup, WAF rules, DDoS and bot attacks. AWS or GCP bill grew and nobody can explain why. Moving from VPS, shared hosting or Replit to a proper setup. Kubernetes and Docker in production, from first cluster to day to day support. CI/CD with GitHub Actions or GitLab, so deploys stop being scary. Monitoring with Prometheus and Grafana that alerts the right person. Linux server administration, security hardening and backups. Self-hosted AI and LLMs on your own GPUs. Ongoing DevOps support for small teams. What I work with: Linux, Docker, Kubernetes, Helm, Terraform, Ansible, AWS, GCP, Azure, GitHub Actions, GitLab CI, ArgoCD, Prometheus, Grafana, Nginx, Cloudflare, Vault, WireGuard, Bash, Python, Go. How I work. First I look at what's actually running. Then I tell you what to fix now and what can wait. Before I touch production I know how to roll it back. I write down what I changed and why, so you're not locked to me. If something can't be fixed fast, I'll say so. Same if I'm not the right person for it. I take one time fixes and long term work. I'm in Ukraine, UTC+2, I have time for full time work now and I usually reply within a few hours. Tell me what's broken and I'll tell you what it takes.