Freelance cloud & AI engineer · taking new projects

Reliable deployments, less manual work, and systems your team can run.

Contributions merged into Kubescape, PyPA build, Terrakube and the Home Assistant docs · Client work for Racoona

TerraformAWSKubernetesAnsibleProxmoxPrometheusGrafanaLangGraphOllamaPythonPodmanKeycloakAlmaLinuxAstroNode.js
8merged open-source PRsAcross 4 projects · Kubescape · Home Assistant · PyPA · Terrakube24hReply timeOverlap with Europe and US mornings5.0★Upwork rating100% job success

A typical setup

Built to run, not just to demo.

Most projects end up with the same backbone: infrastructure defined in code, traffic spread across containers, a managed database and monitoring from day one. You get the setup and the notes to run it yourself.

  • Terraform for repeatable environments
  • Kubernetes or containers behind a load balancer
  • Prometheus & Grafana so problems show up early
gatewayOnline· HTTPS
apiOnline· pod 1
apiOnline· pod 2
postgresOnline· primary
grafana
staging · eu-west-1

Services

Technical work that moves the business forward.

Choose a focused engagement or combine services around one outcome.

Cloud & DevOps setup

A dependable path from infrastructure to production, defined in code and monitored from day one.

  • Terraform
  • AWS
  • Kubernetes
  • CI/CD
  • Prometheus
  • Grafana
  • Keycloak SSO
Typical timeline2–6 weeksGet a quote

AI automation

Useful AI wired into your real infrastructure, workflows and data, with checks so it stays reliable.

  • LangGraph
  • Ollama
  • Python
  • RAG
  • LLM agents
  • AWS
Typical timeline2–6 weeksGet a quote

Business websites

A fast, clear site that loads instantly, works on every screen and turns visits into enquiries.

  • Astro
  • Node.js
  • Responsive UI
  • Contact forms
  • Web3Forms
Typical timeline1–3 weeksGet a quote

Selected work

Proof in the work.

Client delivery and platform engineering grounded in real systems.

Client work

AI-Powered Infrastructure Generation

Problem
Infrastructure requests needed to become reliable, context-aware deployments without repeating manual setup.
Built
A multi-agent system using LLMs and RAG to generate Terraform, deploy to AWS, validate the result and run post-deployment checks.
Result
A connected workflow from infrastructure request through validated AWS deployment.
  • Python
  • LangGraph
  • Ollama
  • RAG
  • AWS
  • Terraform
Racoona AI homepage: “We Invest in Great Hospitality Businesses”

Client work · live site

Racoona

Problem
The business needed a fast static website with a dependable way for visitors to make contact.
Built
Built and maintained the Astro site with contact forms and Web3Forms email delivery.
Result
A lightweight, live website with working contact delivery.
  • Astro
  • Web3Forms
  • Email delivery

Platform engineering

Peracom Internal Developer Platform

Problem
Provisioning and deployment across infrastructure needed a repeatable path with observability.
Built
Automated provisioning and deployment with Proxmox, Terraform, Ansible and Kubernetes, with Prometheus and Grafana monitoring.
Result
A repeatable internal platform workflow with monitoring in place.
  • Proxmox
  • Terraform
  • Ansible
  • Kubernetes
  • Prometheus
  • Grafana

More projects

GloveXcel

IoT rehabilitation glove with sensor processing, BLE communication and web APIs.

  • ESP32
  • BLE
  • React
  • Node.js

Agrimate

Bilingual farmer-to-customer marketplace with crop-care advice and delivery routing.

  • Flutter
  • Node.js
  • Firebase

Open source

Comfortable in other people's codebases.

8 merged pull requests across 4 projects, from Kubescape (12k★) to the Home Assistant docs, each reviewed against the project's own standards.

Show 4 more
All merged PRs on GitHub

How a project runs

A clear path from first conversation to handover.

  1. 01

    Intro call

    We agree on the goal, constraints and what done looks like.

  2. 02

    Proposal and fixed scope

    A written plan with deliverables, timeline and price.

  3. 03

    Build with regular updates

    Short progress updates so you always know where things stand.

  4. 04

    Handover and docs

    Working system, documentation and a walkthrough for your team.

  • Response within 24 hours
  • Based in Sri Lanka (UTC+5:30)
  • Overlap with Europe and US mornings
Chanuth Abeynayake

About

Chanuth Abeynayake

Freelance cloud & AI engineer · Sri Lanka

I'm Chanuth, a freelance cloud and AI engineer. I set up infrastructure your team can run on its own, with Terraform, Kubernetes and monitoring from day one, and I build AI automation that plugs into real workflows. I also contribute fixes and features to open-source projects such as Kubescape.

Computer Engineering undergraduate at the University of Peradeniya. Team lead in three hackathon finals: SLASSCOM Xcelerate, SLIIT CodeFest and SLT CodeBlast (LinkedIn ↗).

Contact

Have a system to improve?

Tell me what you need, and I’ll reply with the next sensible step.

hello@chanuth.dev