Available for new projects

Building reliable infrastructure and tools people enjoy using.

DevOps engineer at an AI company. 4+ years running Kubernetes, self-hosted LLM inference, and the boring machinery that keeps products online. Fixed scope, fixed price, first intro call free.

DevOps engineer · Vilnius, Lithuania · Evenings & weekends

View My Work
pi-4 · ssh

martyn2s@pi-4:~$ kubectl get pods -A

NAMESPACE     NAME                    READY  STATUS

llm           vllm-7b-6d9f4b-x2k      1/1    Running

monitoring    grafana-7c8b9d-q4        1/1    Running

apps          purrdo-web-5f7c-z9      1/1    Running

martyn2s@pi-4:~$

grafana · homelab overview

CPU

42%

MEM

68%

GPU

91%

What I can build for you

Fixed scope, fixed price. You know what you get before we start.

01

Kubernetes audit & rescue

Your cluster is unstable, expensive, or nobody fully understands it anymore. Full health check: workloads, security posture, networking, and where the money is leaking.

Written report · prioritized fix list · scope agreed upfront

02

Self-hosted LLM infrastructure

Private AI inference on your hardware or in a private cloud, built on vLLM. OpenAI-compatible API, GPU scheduling, monitoring, autoscaling. Your data never leaves your perimeter.

Production deployment · documentation · scope agreed upfront

03

Production readiness

You have a product that works and infrastructure held together with hope. I set up CI/CD, Cloudflare, error tracking, alerting, and backups. The boring things that keep you online.

Pipelines + observability · runbooks · scope agreed upfront

04

Infrastructure retainer

Ongoing care after the setup: patching, monitoring, small improvements, all in a fixed block of hours each month. A part-time DevOps engineer, without the hiring process.

Monthly · fixed hours · planned work only

Selected Work

Products I designed, built, and run in production myself.

Like what you see?

Tell me about your setup. The first intro call is free.

Get in Touch

How I work

No account managers, no ticket queues. You talk to the person doing the work.

  1. 01

    You describe the problem

    A form message or a short call. I ask about your stack, what's breaking, and what 'done' looks like for you. No commitment on either side.

  2. 02

    You get a written scope

    What I'll deliver, what it costs, what I need from you, all before any work starts. Fixed price for fixed scope. If I'm not the right fit, I say so and point you elsewhere.

  3. 03

    I build, you see progress

    Work happens in the evenings and weekends, in small visible increments. You get access to the repo and environment as things take shape. No big-bang reveal.

  4. 04

    Handover you can own

    Documentation, runbooks, and a walkthrough call. The goal is that your team can run it without me. I'm around if you want more later, not because you're locked in.

Questions, answered

The things people usually ask before the first call.

What is the free intro call?+

30 minutes, no slides and no pitch. You describe your setup and what's hurting, I ask questions, and we both figure out if there's a fit. After it you get a written scope, or an honest 'I'm not the right person for this.'

How quickly can we start?+

I take on a small number of projects at a time and work evenings and weekends. The intro call is free. If my queue is full or I'm not the right fit for your problem, I'll tell you straight and point you elsewhere.

How does fixed pricing work?+

Scope first: before any work starts, you get a written definition of what I'll deliver and what it costs. That price doesn't change. If the requirements change mid-way, we write a new scope for the extra work. No surprise invoices.

Who owns the code and infrastructure?+

You do. You get access to the repo and environment from day one, and documentation and runbooks at handover. No lock-in. The goal is that your team can run everything without me.

Do you work with companies outside Lithuania?+

Yes. I work in English and Lithuanian, remote-first. Contracts and invoicing are straightforward within the EU.

What if something breaks after handover?+

Handover includes documentation and a walkthrough call so your team can handle day-to-day issues. If you want ongoing care (patching, monitoring, small improvements), that's the monthly retainer: planned work in a fixed block of hours, not a 24/7 on-call contract.

About

Location
Vilnius, Lithuania
Experience
4+ years in DevOps
Education
BSc IT, MSc Strategic ISM, Vilnius University
Languages
Lithuanian, English

I'm a DevOps engineer at an AI company, based in Vilnius, with 4+ years in the field. Day to day I run Kubernetes, GPU workloads, and self-hosted model inference with vLLM, with Cloudflare, Sentry, Teleport, Backstage, and GitLab around them. Terraform, Ansible, and Bash are how I get things done. Beyond the technical work, I write and triage Jira tickets, groom the backlog, assign work, and build the infrastructure processes the team runs on.

Outside of work I build and run my own projects, like Purrdo, on a self-hosted Kubernetes homelab. The thread through all of it: infrastructure deserves the same design care as the product it runs.

Contact
Available for new projects

Let's make your infrastructure boring. In a good way.

The form lands straight in my inbox, and a few sentences about your setup is enough. I take on a small number of projects at a time, and the first 30-minute intro call is free.

Your message goes straight to my inbox. No tracking, no marketing lists. Privacy notice