Infrastructure consulting
Cut the bill. Raise the SLA. Harden the pipeline.
I help companies reduce infrastructure cost by up to 90%, raise availability up to a 99.96% SLA, and build security into delivery instead of bolting it on after an audit.
Hamidreza Shafiee Senior DevOps, DevSecOps and security engineer · Dubai
I don’t settle for the first option. I look for, or build, a better one.
About
A senior engineer for the decisions that show up on the bill, the incident, and the audit.
I am based in Dubai and work remotely with companies in the United States, Canada, the UAE, the United Kingdom, Germany, China, and Russia.
Since July 2024 I have been DevOps Team Lead and DevOps Manager, remotely, for a company that specializes in LLM and MLOps on Azure and AWS. The employer is not public. Before that I led the DevOps team at Hamrahe Aval (MCI), a major regional telecom. The path here is more than ten years, from network and Linux engineering to leading DevOps and DevSecOps teams.
Services
Four ways I take work off your critical path.
Engagements are scoped to the estate in front of us. The percentages on this page are the ceiling of results I have delivered, not a quote for an environment I have not seen.
Infrastructure cost reduction
Bills hide waste in idle clusters, oversized nodes, cross-zone traffic, and tools that no longer earn their seat. I trace the spend to the workloads that cause it, remove what is not earning its keep, and redesign the platform so the saving survives the next quarter. The ceiling on this work is a reduction of up to 90%.
Uptime and reliability
The gap between 99.9% and 99.96% is not a brighter dashboard. It is fewer failure domains, safer releases, and an on-call path that reaches the person who can fix the fault. I have raised availability up to a 99.96% SLA by changing how the system fails and how the team ships.
Security and DevSecOps
Security that shows up after an audit is already late. I put identity, secrets, image and dependency gates, network policy, and reviewable infrastructure into the path code takes to production. The control lives in the pipeline. It is not a document someone hopes was followed.
Kubernetes and GitOps platforms
When the cluster, the GitOps controller, and the infrastructure code are what your engineers touch every day, they have to be boring to operate. I design and build Kubernetes platforms with GitOps (Argo CD or Flux), Terraform, and a promotion path your team can run after I leave.
Tools I actually use
- Kubernetes
- GitOps (Argo CD and Flux)
- Terraform and infrastructure as code
- HashiCorp Vault
- Ansible
- Docker
- CI/CD security
- AWS
- Azure
- Linux
- Nginx
- Networking
- MLOps and LLM infrastructure
Where this work shows up
Industries, not client names.
I do not publish who I have worked for as a client. These are the kinds of environments the work comes from, and the kind of problem each one usually brings.
A major North American telecom operator
The problem
A production estate that grew faster than the operating model. Capacity, change risk, and security controls no longer matched the scale of the network.
The approach
Find where the money and the failures actually sit. Right-size what is idle, put delivery under GitOps, and make infrastructure changes reviewable before they reach production.
Large telecom operators and network vendors
The problem
Many networks and many tools, with little shared practice for how a change is built, reviewed, and rolled forward.
The approach
A common infrastructure-as-code baseline, secrets in Vault rather than in chat, and a pipeline that rejects a change that cannot explain itself.
Power plants and energy
The problem
Systems that cannot be casually restarted, and appliances that vendors leave unpatched because “it is the plant”.
The approach
Separate the change path for plant-adjacent systems from the one used for ordinary IT. Patch, backup, and page on the failure that matters, not on every blip.
Payments and fintech
The problem
Pressure to ship faster sitting next to pressure to prove, later, that someone was in control of identity, secrets, and change.
The approach
Put identity, secrets, and policy in the pipeline so the control is evidence a reviewer can open, not a slide produced the week before the audit.
A crypto exchange
The problem
A high rate of change, an attractive target, and downtime that customers treat as a trust failure rather than a technical one.
The approach
Limit blast radius, keep key material in a system built for it, isolate environments, and make the release path safer before anyone talks about the bill.
International media and news
The problem
Traffic that spikes with the news, on a stack that grew by addition. Publish night should not need a hero.
The approach
Capacity that can rise for the spike and fall after it, plus a delivery path the desk can use without a platform engineer in the room.
Cloud and SaaS platforms
The problem
The product is the platform, so the infrastructure bill is a product problem, and a noisy neighbour is a customer problem.
The approach
Unit cost, tenancy boundaries, GitOps promotion between environments, and security checks on every merge rather than every quarter.
Government, public sector, and universities
The problem
Long-lived systems, shared responsibility, and procurement rules that punish a design only one person understands.
The approach
Documented infrastructure, least privilege, and a handover a team that was not in the room can still operate.
Where the work is cost or availability, the ceiling I have delivered is an infrastructure bill cut by up to 90% and an SLA raised up to 99.96%. Those figures are not a measurement of any single engagement, and they are not a promise before an assessment.
Engagement
Assessment, plan, implementation, then support.
You should know what I found, what I intend to change, and what you will own when I step back.
-
01
Assessment
I read the estate, the bill, the recent incidents, and the pipeline. You get a short written view of what is expensive, what is fragile, and what is unsafe.
-
02
Plan
A sequenced plan: what changes, in what order, what it costs to do, what it risks, and what “done” means. No twelve-month transformation theatre.
-
03
Implementation
I build it with your team, in your repositories, with changes you can review. The result is a system you own, not a black box that needs me to stay.
-
04
Ongoing support
Handover and documentation, and a retainer if you want a senior engineer on the decisions that follow: the next incident, the next bill, the next audit.
Experience
From networks and Linux to leading the platform.
Employers are named. Clients are not. The current role is remote; the employer is not public.
-
DevOps Team Lead / DevOps Manager
DevOps lead for a company that specializes in LLM and MLOps on Azure and AWS. The employer is not named. Secure CI/CD, Terraform, and Kubernetes for machine-learning workloads, with security through the lifecycle.
-
DevOps Team Lead
Led the DevOps team at a major regional telecom. GitOps, Kubernetes, HashiCorp Vault, and Terraform for secure infrastructure and vulnerability remediation.
-
DevOps Engineer
Security automation with Ansible for configuration, patching, and vulnerability remediation, plus secure remote access and monitoring.
-
DevOps Engineer
DevOps for a cloud gaming platform: GitOps, Kubernetes, and security in the CI/CD pipeline, including network segmentation and access control.
-
DevOps Engineer
Software delivery with an emphasis on security, scalability, and technical documentation.
-
Network and Security Manager
Led network security: protocols, threat mitigation, and the security practices used across the company.
-
Senior Network Engineer
Network performance, infrastructure, and security incidents.
-
Senior Network Engineer
Network and wireless services, including LAN, WAN, and internet systems.
Education
Certifications
Contact
Book a call.
Tell me what is expensive, what is fragile, or what an auditor will ask about next. I answer personally. There is no form.