Mohamed ARKID
DevOps & Cloud Consultant
Overview
About
I help companies stop wasting money on broken infrastructure and slow deployments. Most of my work involves taking messy, manually-managed AWS environments and turning them into automated, self-healing systems that actually work at 3 AM without waking anyone up.
Day to day, that means building Kubernetes clusters, writing Terraform, setting up CI/CD pipelines that don't break on Fridays, and finding the $15K/month NAT Gateway charges nobody noticed. I've done this across startups, fintech, and e-commerce.
If your team is afraid to deploy, your AWS bill keeps climbing, or your "monitoring" is someone checking Slack for complaints, we should talk.
Stack
Results
Real outcomes from real engagements. Names withheld under NDA.
Series B SaaS Startup
Challenge: Monolithic Node.js application deployed via manual SSH scripts. Bi-weekly releases with 2+ hours of downtime per deployment.
Action: Containerized the application with Docker, deployed to EKS with Terraform, and implemented a full GitOps pipeline using ArgoCD with Canary rollouts.
Result: Deployment frequency: bi-weekly to 8x/day. Downtime: 0.
FinTech Company (100+ Engineers)
Challenge: AWS bill growing 15% month-over-month with no visibility into cost drivers. NAT Gateway charges alone were $12,000/month.
Action: Implemented VPC Endpoints to eliminate NAT Gateway data transfer costs, migrated stateless workloads to Spot Instances, and integrated Infracost into CI/CD.
Result: Reduced monthly AWS spend by 38% ($45K/mo saved).
E-Commerce Platform
Challenge: No monitoring beyond CloudWatch basic metrics. Average incident detection time was 45 minutes, often discovered by customers before engineers.
Action: Deployed a full observability stack (Prometheus, Grafana, Loki) with Golden Signals alerting and automated Slack escalation via n8n.
Result: Mean Time To Detection dropped from 45 minutes to under 2 minutes.
Experience
- Infrastructure Architecture: Architected and managed AWS cloud infrastructure for a high-volume EDI validation platform, ensuring sub-second API latency for thousands of daily payloads.
- Kubernetes: Built a production-grade Kubernetes cluster from scratch on VMware for internal microservices, eliminating deployment bottlenecks and increasing release velocity by 300%.
- Zero-Touch Automation: Replaced manual runbooks with automated n8n workflows for CI/CD and monitoring, saving 15+ hours of engineering toil per week.
- Backend Scaling: Deployed and secured FastAPI backend services on AWS, maintaining strict compliance for sensitive supply chain data.
- AWS
- Kubernetes
- Docker
- VMware
- FastAPI
- n8n
- CI/CD
- Cloud Security
- Monitoring
- GitLab CI/CD
- Docker
- Terraform
- AWS
- Kubernetes
- Prometheus
- Cloud Infrastructure
- Automation
AJICOD
- Java
- Spring Boot
- PostgreSQL
- REST API
- Authentication
- Docker
Education
- Embedded AI
- Software Engineering
- Systems Design
- Java
- Python
- PostgreSQL
- Software Engineering
Latest Posts
View all
Your AWS Bill is 30% Too High: The Architect's Guide to Slashing Cloud Costs
Stop wasting money on over-provisioned infrastructure. Learn the exact engineering strategies—from Spot Instances to Karpenter—that I use to instantly cut AWS and Kubernetes bills by 30% without sacrificing performance.
Kubernetes on Bare Metal: Why It's Harder Than You Think (And Why It's Worth It)
Most teams run Kubernetes on managed cloud services and never think twice. But what happens when you strip away the safety net — no load balancer API, no CSI magic, no managed control plane? This is the reality of bare-metal Kubernetes and why mastering it makes you a better engineer.
Stop Grepping Logs: Building an Observability Stack That Actually Tells You What's Broken
If your debugging process starts with 'grep' and ends with 'I have no idea what happened,' your monitoring is fundamentally broken. Here is how to build a Prometheus, Grafana, and Loki stack that pinpoints failures in seconds.