CompagAI

Cloud, platform & AI infrastructure engineering

Build infrastructure that just works.

CompagAI designs, automates and runs Azure, Kubernetes and AI infrastructure — so your engineers ship product instead of fighting platforms.

Start a project See our open-source work

How we ship: code to production on AKS

Deployment pipeline to Azure Kubernetes Service A pull request in GitHub or GitLab triggers a CI/CD pipeline that builds, tests and scans a container image, pushes it to Azure Container Registry, and performs a rolling update on an AKS cluster with an application node pool and a GPU inference node pool. Terraform provisions the registry and the cluster. User traffic enters through a TLS ingress, and Azure Monitor collects metrics and logs. 00 provision 01 commit 02 verify 03 publish 04 deploy 05 observe Source GitHub · GitLab PR #142 → main CI/CD pipeline build test scan image + SBOM Registry Azure ACR app:v42 pushed Users · HTTPS Ingress · TLS · WAF app pool api api web web HPA · 4–12 pods gpu pool inference inference GPU · scale to zero AKS cluster v41 running rolling update… v42 healthy ✓ Azure Monitor metrics · logs · alerts Terraform network · ACR · AKS · identity
Every change goes through review, automated tests and a security scan before a zero-downtime rolling update — on infrastructure that is itself defined in code.
Ex-Microsoft Azure engineeringYears building and operating Azure platform services from the inside.
14+ years in DevOpsFrom classic release pipelines to GPU node pools.
Microsoft certifiedAZ-400 DevOps Engineer Expert · AZ-305 Solutions Architect Expert
Top Rated Plus on UpworkA consistent delivery record with international clients.

Engineering across the full delivery path

Hands-on work from the landing zone to the GPU node pool — designed, built in code, and handed over so your team can run it.

Cloud & platform engineering

Azure cloud

Secure, well-governed Azure environments built to scale.

  • Landing zones & governance
  • Networking & private endpoints
  • Identity, RBAC & Key Vault
  • Cost management

Terraform & IaC

Every environment reproducible from code.

  • Reusable modules
  • Remote state & locking
  • Multi-environment workflows
  • Drift detection & policy

CI/CD pipelines

Pipelines that move code safely from commit to production.

  • GitHub Actions
  • GitLab CI/CD
  • Azure DevOps YAML
  • Approvals, gates & rollbacks

Kubernetes & AKS

Container platforms your teams can trust in production.

  • AKS design & hardening
  • Helm & GitOps
  • Autoscaling & workload identity
  • Observability

AI & compute infrastructure

AI infrastructure

The platform layer underneath your AI applications.

  • Azure OpenAI & AI Foundry
  • AI gateways, quotas & cost control
  • RAG pipelines & vector search
  • GPU node pools for inference

HPC & Slurm

Batch and high-performance compute that scales on demand.

  • Slurm clusters on Azure
  • Autoscaling partitions
  • Shared storage & job accounting
  • Licence-aware scheduling

Python & automation

Tooling that removes the toil from operations.

  • FastAPI services & internal APIs
  • Infrastructure automation
  • Log & telemetry processing
  • CLI tools & integrations

Ongoing DevOps & SRE

Engineering support after go-live.

  • Production support
  • Incident root-cause analysis
  • Performance & cost tuning
  • Runbooks & documentation
Tools we work with AzureAKSTerraform GitHub ActionsGitLab CIAzure DevOps DockerHelmSlurm PythonFastAPIPrometheus GrafanaAzure OpenAI

Building in public

We're publishing reference implementations of the platforms we build — real, deployable code you can read before you work with us.

GitHub GitLab
aks-platform

Production-ready AKS in Terraform — networking, ACR, workload identity, autoscaling and monitoring, with CI/CD included.

TerraformAKSAzure
In progress
ai-gateway

A FastAPI gateway for Azure OpenAI with per-team token quotas, rate limiting, streaming and cost tracking.

PythonFastAPIAzure OpenAI
In progress
slurm-on-azure

A Slurm HPC cluster on Azure with autoscaling compute partitions, shared storage and job accounting.

SlurmTerraformHPC
Planned
pipeline-templates

Reusable, secure CI/CD templates for GitHub Actions, GitLab CI and Azure DevOps — build, scan, deploy and roll back.

GitLab CIGitHub ActionsAzure DevOps
Planned
rag-infrastructure

Reference RAG infrastructure — document ingestion, embeddings, Azure AI Search and scheduled index refresh.

PythonAI SearchTerraform
Planned

Follow the repositories to be notified as each one ships.

Simple, transparent, engineering-first

The same four steps on every engagement, whether it's a two-week pipeline migration or a full platform build.

01

Understand

We map your applications, infrastructure, constraints and how your team delivers today.

02

Design

We propose an architecture built for reliability, security and cost — and agree it with you before building.

03

Automate

Everything goes into code: Terraform, pipelines, policies and runbooks. Nothing is click-configured.

04

Hand over

We test, document and walk your team through it, so you can operate it without us.

Ways to work with us

Pick the model that fits the problem. Every engagement starts with a short scoping call.

Fixed-scope project

A defined outcome delivered against an agreed plan and timeline.

Landing zone · AKS platform · pipeline migration

Monthly retainer

Dedicated engineering capacity for ongoing platform work, improvements and support.

Platform ops · SRE · continuous improvement

Review & advisory

An independent review of your cloud, pipelines or AI platform, with a prioritised action plan.

Security · cost · reliability · architecture

Have an infrastructure challenge?

Tell us what you're building, migrating, automating or fixing. We'll come back with questions, an approach and a clear next step.

Email us directly

Start a project Hire us on Upwork