Serhii KrepetsDevOps / Platform Engineer

I design and build the systems that get code from a commit into production, and keep it running once it's there — CI/CD pipelines, infrastructure as code, and monitoring.

Full‑stack background (2023–) + DevOps/Platform focus (2025–) + AI‑assisted engineering workflow.

title	portfolio — s. krepets
discipline	devops / platform / full‑stack
location	halifax, ns, ca
sheet	1 of 7
scale	n.t.s.
rev	2026.07
status	available for work

SH‑02 / COMPONENTS

What I Do

Infrastructure

CI/CD pipelines, containerized deployments, infrastructure as code, and cloud environments — the systems that keep applications running after the code is written.

Application Layer

Next.js applications, React frontends, and Node.js / NestJS backends — plus commercial PHP / WordPress systems. Knowing what the app needs shapes how I build the infrastructure under it.

AI‑Assisted Workflow

AI tools built into the process — for research, debugging, documentation, and infrastructure planning. A multiplier on judgment, not a replacement for it.

SH‑03 / NOTES

Specification Notes

NOTE 1Who I Am

DevOps / Platform engineer. Infrastructure, CI/CD, and deployment are what I own — not a few skills bolted onto a developer title. I started on the application side, which is exactly why I don't treat what happens after the code ships as someone else's problem.

NOTE 2Where I'm Headed

Full‑stack development was 2023–2025 — the foundation, not the pitch. DevOps / Platform is the track built since, and it's the one still growing. On a longer horizon, AI is closing enough of the distance between idea, design, and shipped product that owning that whole path solo stops looking unrealistic — a direction worth watching, not a plan for right now.

full‑stack — foundationdevops / platform — nowplatform engineer — target

NOTE 3What I Bring

Hire me for DevOps / Platform work — the edge is being able to read and reason about the application running on that infrastructure, not offering "a bit of everything." Infrastructure decisions come from real application experience, not guesswork. The team matters as much as the stack: I look for a company worth investing in, and once I'm in, I bring ideas instead of waiting for a ticket.

SH‑04 / ARCHITECTURE

Deployment Architecture

full history on github→

shipped — client & freelance work

client-vps-hostinginfra / hosting2023—

problem: Clients had a finished site and no idea where or how to host it — and no budget for a managed platform.
solution: Provisioned Ubuntu VPS instances on Hetzner Cloud, configured Nginx as a reverse proxy with automatic SSL via Let's Encrypt, set up DNS, and locked each server down with a basic firewall.
stack: Hetzner Cloud · Ubuntu · Nginx · Let's Encrypt · UFW
result: Several client projects went from "no idea how to launch this" to a production server that just kept running.

github-push-deployinfra / ci‑cd2024—

problem: Releasing an update meant SSH-ing into the server, pulling the repo, and restarting the process by hand — easy to forget a step.
solution: Set up GitHub Actions / webhook-triggered deploy scripts so a push to main pulls the latest code, installs dependencies, builds, and restarts the app — no manual steps.
stack: GitHub Actions · Webhooks · SSH deploy keys · systemd / PM2
result: "Push to deploy" replaced a manual checklist that depended on remembering the right order of commands.

aws-amplify-hostinginfra / cloud2024

problem: A project needed hosting with automatic builds and per-branch previews, without managing a server.
solution: Connected the repo to AWS Amplify, configured build settings and environment variables, and set up branch-based preview deployments alongside the production build.
stack: AWS Amplify · Next.js · Managed CI/CD
result: Hands-on experience with a managed cloud deploy pipeline outside the usual Vercel/Netlify defaults.

coolify-selfhosted-paasinfra / platform2025—

problem: Wanted the "push and it's live" convenience of a PaaS like Vercel or Heroku, but running on infrastructure I actually control.
solution: Deployed Coolify on a Hetzner VPS as a self-hosted PaaS — apps, databases, and preview environments all run through its UI on top of Docker.
stack: Coolify · Docker · Hetzner Cloud
result: A working self-hosted alternative to a managed PaaS — and the direct inspiration for the platform below.

building now — my own platform

k8s-blue-green-platforminfra / kubernetesin progress

problem: A single Docker host can't do a real zero-downtime release — a bad deploy still means a visible blip or a manual rollback.
plan: Standing up a Kubernetes cluster with a blue/green rollout pipeline: the new version deploys alongside the old one, passes health checks, then traffic switches over — with an instant rollback if it doesn't.
stack: Kubernetes · Helm · Ingress‑nginx · GitHub Actions
status: In progress — building this while job‑hunting, not a finished product yet.

easy-deploy-platforminfra / platformplanned

problem: Every new project still means wiring up CI/CD and hosting by hand — the exact repetitive work the platform above should remove.
plan: A small self-serve platform on top of the Kubernetes cluster: point it at a git repo through a UI, and it builds, containerizes, and deploys the app to my own server automatically — with logs and rollback built in.
stack: Kubernetes · Docker · Next.js (control panel) · GitHub API
status: Planned — next up once the cluster and blue/green pipeline are in place.

nextjs-supabase-appplatform / reference‑appplanned

problem: The platform above needs a real workload to prove it works — not just a "hello world" container.
plan: Building a Next.js + Supabase app (auth, Postgres, storage) as both a standalone project and the first real deploy target for the platform.
stack: Next.js · Supabase · PostgreSQL
status: Planned — the app that puts the platform through its paces.

SH‑05 / BUILD LOG

Experience

Developer since 2023 — full‑stack first, with a DevOps / platform focus that's newer, but growing fast.

full‑stack / frontend

devops / platform

Full‑Stack / Frontend Development

stable · since 2023

Building and shipping production web applications since 2023 — Next.js / React frontends and Node.js backends, across freelance work and full‑time roles, from pixel‑perfect implementation to deployment.

DevOps / Platform Engineering

in progress · since 2025

Took on DevOps‑focused freelance work in 2025 — CI/CD pipelines, deployment automation, and infrastructure for client projects via Upwork and Fiverr. The newest track, and the one growing the fastest.

SH‑06 / CHANGE REQUEST

Contact

available for work — Halifax, Canada

Open to DevOps / Platform / Cloud / Infrastructure work — remote or hybrid.