Mohamed ARKID logo

Command Palette

Search for a command to run...

Command Palette

Search for a command to run...

Illustration: an AWS bill where idle non-production resources, NAT Gateway data charges, oversized EC2 instances and orphaned EBS volumes are removed, leaving the same workloads on a smaller bill.

Portrait of Mohamed ARKID
Resume

I cut AWS bills for startups

Production Work & Case Studies

Flagship Case Studies

Real production architectures built for startups and fast-growing teams. Measured in slashed AWS bills, zero-downtime releases, and high-reliability platforms.

AWS FinOps & Cloud Infrastructure

Automated infrastructure-as-code solution deploying resilient AWS multi-AZ architectures with strict least-privilege IAM policies, autoscaling groups, and proactive cost governance.

-40% AWS Waste100% Terraform IaCZero Drift
Read Case StudyGitHub Repo
terraform-aws-finops.tf
ACTIVE • US-EAST-1 • -40% SPEND
terraform apply -auto-approve
Initializing provider plugins: aws ~> 5.40, kubernetes ~> 2.26
+ module.vpc.aws_subnet.private_subnets [us-east-1a, 1b]
+ module.eks.aws_autoscaling_group.worker_nodes (min: 3, max: 12)
+ aws_cur_budget_action: auto-shutdown idle non-prod workloads
Apply complete! Resources: 38 added, 0 changed, 0 destroyed.
⚡ Impact: AWS monthly waste slashed from $14,200 to $8,500 (-40.1%)

Zero-Downtime Kubernetes & GitOps

Production-grade container platform built on Kubernetes and managed entirely through GitOps with ArgoCD. Features automated horizontal pod autoscaling (HPA) and automated canary rollbacks.

99.99% AvailabilityZero Deploy DowntimeArgoCD GitOps
Read Case StudyGitHub Repo
k8s-gitops-rollout.yaml
K8S • ARGOCD • 3/3 HEALTHY
kubectl get pods -n production -l tier=api
NAME READY STATUS RESTARTS AGE
api-gateway-7b9c-4kx82 1/1 Running 0 42d
api-gateway-7b9c-9mj12 1/1 Running 0 42d
api-gateway-7b9c-x10qp 1/1 Running 0 42d
✓ ArgoCD Sync: Synced (Healthy) | Rollout: 100% Success
⚡ Traffic: 0 dropped packets during rolling deployment | HPA active

Observability & Four Golden Signals

Comprehensive 9-service telemetry engine capturing Latency, Traffic, Errors, and Saturation with Prometheus, Grafana, and Loki. Features automated anomaly alerting thresholds and actionable runbooks.

< 28ms Latency9-Service StackAlertManager Runbooks
Read Case StudyGitHub Repo
telemetry-golden-signals.promql
TELEMETRY • NOMINAL • 4 SIGNALS
promtool check rules alertmanager.yml
Checking alerts and recording rule syntax...
✓ Latency p99: 28.4ms (SLA threshold < 120ms)
✓ HTTP Error Rate: 0.002% (2 failures / 100k requests)
✓ Cluster Saturation: CPU 32% | Memory 54% (Nominal)
✓ AlertManager: 0 Firing | 4 Runbooks Verified
⚡ Observability: Real-time telemetry across all microservices

Overview

Independent DevOps & Cloud Consultant

Remote — Morocco (UTC+1)

[Email protected]

moarkid.com

Social Links

LinkedIn

GitHub

Production Competencies

Technology Stack & Systems Matrix

LAYER 01/

Cloud Infrastructure

Enterprise Multi-AZ & Cloud Economics

LAYER 02/

Containers & Orchestration

Zero-Downtime Cluster Management

LAYER 03/

IaC & GitOps Automation

Immutable Delivery Pipelines & Zero Drift

LAYER 04/

Observability & Golden Signals

Proactive Telemetry & Anomaly Alerting

LAYER 05/

Systems & Core Languages

High-Throughput Linux Kernels & APIs

Projects(4)

Construction-site management for Moroccan BTP companies on Odoo 19 Community, with no enterprise licence.

  • Field invoice intake: site managers send a photo or PDF to a Telegram bot; n8n runs it through self-hosted OCR (or a vision API with automatic fallback) and creates a draft vendor bill in Odoo.
  • Posting guard: every OCR bill lands as a draft, and low-confidence reads are flagged for an accountant to check against the original.
  • Moroccan business rules: progressive billing (situations de travaux), guarantee retention with the 7% public-contract cap, a real-time margin view per site, and a printable A4 situation with tripartite signatures.
  • Operations: Docker Compose, deploy on green CI with a database copy before each release and automatic rollback, scripted database and filestore backups.
Read the case study
  • Odoo 19
  • Python
  • n8n
  • Tesseract OCR
  • PostgreSQL
  • Docker

A live tool that explains why a domain's email fails authentication, and decodes bounce messages into a cause and a fix.

  • Checks SPF, DKIM, DMARC and MX, follows every SPF include, counts DNS lookups against the limit of 10, and hands back the exact record to publish.
  • Fast: independent lookups run in parallel, which cut a check from 3–4 seconds to under half a second in replay tests.
  • Production-run: CLI, HTTP API and web page, 124 tests, per-client rate limiting, deploy on push to main, and an hourly uptime and certificate check from outside.
  • Proven on a real domain: used to set up sending for moarkid.com, which now passes SPF, DKIM and DMARC in Gmail and scores 10/10 on mail-tester.
Read the case studyWatch the demo
  • Python
  • FastAPI
  • DNS
  • Docker
  • Caddy
  • GitHub Actions

Tools & Open Source

GitHub

AWS Cost Waste Calculator

Free tool

Free interactive tool — estimate how much of your AWS bill is recoverable waste.

aws-ec2-reaper

GitHub

Lambda + Terraform that stops tagged non-prod EC2, RDS and ECS outside working hours and starts them again in the morning.

aws-cur-analyzer

GitHub

Parse AWS Cost & Usage Reports with Python to surface your biggest per-resource cost offenders.

aws-vpc-endpoints-terraform

GitHub

Terraform module that routes S3, DynamoDB and other AWS service traffic through VPC endpoints instead of the NAT Gateway.

Where I've Worked

Production track record across cloud infrastructure, Kubernetes platforms, and automated delivery pipelines.

Resume (PDF)

DevOps Engineer @ PROCIA

01.2026 — 05.2026•Internship
  • Deployed and operated containerized AI automation services on Kubernetes (Deployments, Services, Ingress, ConfigMaps/Secrets).
  • Packaged services with Docker and managed releases through CI/CD pipelines.
  • Tuned resource requests/limits and autoscaling so automation jobs ran reliably under variable load.
  • Monitored cluster health and troubleshot pod, networking and scheduling issues in production.
  • Worked with the development team to move new AI features from build to deployment.
KubernetesDockerCI/CDLinuxPython

Education

Academic qualifications in embedded systems, digital services, and software engineering.

Master's Degree, Embedded Systems and Digital Services @ Université Ibn Zohr

2023 — 07.2025•Master of Science•Agadir, Morocco
  • Embedded Systems & Architecture: Advanced study in real-time embedded systems, kernel modules, microcontrollers, and hardware-software integration.
  • Digital Services & Networks: Architected distributed digital services, high-throughput network protocols, and resilient IoT edge automation.
  • Engineering Delivery: Applied disciplined software engineering, containerized services with Docker, and automated deployment pipelines to technical systems.
Embedded SystemsDigital ServicesLinuxC/C++PythonDockerSoftware Engineering

Writing

All posts

AWS Bill Too High? 7 Fixes With Copy-Paste Terraform

Eliminate AWS NAT Gateway Overcharges: Terraform Guide

AWS Compute Cost Playbook: Savings Plans, RIs & Spot

Free PDF checklist

The 10-point AWS cost checklist

The checks I run first on an AWS bill, each with the AWS CLI command that finds the waste and what to do about it. You can also read it on the site.

No spam. Unsubscribe any time.