DevOps and cloud infrastructure that is faster to ship, steadier to run and cheaper to keep

Who this is for#

Startups whose deployment is a checklist in a doc, teams whose cloud bill grew faster than their revenue, AI products that need inference services and GPU-aware scheduling in production, and companies that want their infrastructure written down as code before the one person who understands it leaves.

What you get#

• CI/CD pipelines (GitHub Actions, GitLab, Travis) that test, build, scan and deploy on every merge, with preview environments where they help.
• Containers and orchestration: Docker images done right, Kubernetes (AKS, EKS, GKE) with autoscaling, health checks, rolling deploys and rollbacks — or a simpler platform when that's the honest answer.
• Infrastructure as code with Terraform, so environments are reproducible and reviewed like any other change.
• Observability: Grafana dashboards, logs and traces, alerts that page for the right things, and load tests (JMeter) before traffic surprises you.
• Cost and performance tuning: right-sizing, autoscaling policies, caching, database tuning, and inference cost work for AI services.
• A runbook your team can follow at 3 a.m.

How I work#

1. Audit. I read the current setup — pipelines, hosting, secrets, backups, monitoring, bill — and give you a written list of risks and quick wins, ranked.
2. Fix the sharp edges first. Backups, secrets, a repeatable deploy. Usually the first week.
3. Build the platform. Infrastructure as code, pipelines, environments, observability.
4. Hand over. Documentation, a runbook, and a session with your team so it stays boring after I leave.

Proof#

• I run the production infrastructure of a real-time voice AI platform on Azure — Kubernetes, PostgreSQL, monitoring — serving dental offices across the United States 24/7.
• At Syntheia.io I led backend and MLOps work that improved system performance by 30% and cut inference time by 20%; a microservice redesign lifted stability by 25%, code coverage by 50%, and brought a slow query from 2 seconds to 10 milliseconds.
• Udagram: refactored a monolithic Node.js app into microservices with a complete CI/CD pipeline (Travis CI), an NGINX API gateway and deployment to AWS EKS.
• I've deployed ML backends to several cloud providers and to on-premise servers, including LLM and generative-AI services.
• I also teach this: my Udacity webinars cover Docker, Kubernetes and AWS deployment, and I'm a 5-star rated mentor there.

Stack I typically use#

Kubernetes (AKS, EKS, GKE) · Docker · Terraform · GitHub Actions, GitLab CI, Travis · Azure, AWS, GCP · NGINX · PostgreSQL, Redis, Kafka · Grafana, Prometheus, JMeter · Linux.

Where and how I work#

Remote from Cairo (UTC+2), with clients in the United States, Europe and the Middle East. Everything I change is in code and reviewed, so you can see exactly what happened and why.