DevOps engineer and programmer
I build the pipelines that take code from a commit to production, and I still write the code that goes through them. I've been a DevOps engineer at Microsoft since 2023, after nine years building websites and web applications.
Based in Israel and open to relocation.
Prefer a shell? Press ` or .
career / main
Queued
I help teams ship software more often and with less risk. These are the areas where I can take ownership from day one.
Pipelines that build, test and deploy every change the same way, from the first commit to production, with GitOps delivery through ArgoCD.
Jenkins, GitHub Actions, GitLab CI, ArgoCD
Cloud environments described in Terraform or CloudFormation, so staging and production match and can be rebuilt on demand.
Terraform, CloudFormation, AWS CLI, EC2, S3, IAM, VPC
Servers configured by Ansible, Puppet or Chef instead of by hand, so they don't drift apart over time.
Ansible, Puppet, Chef, Ruby, Python
Planned releases with clear steps, sign-off from dev, QA and ops, and a rollback plan ready before anything ships.
Jenkins, runbooks, rollback plans
Ubuntu and Debian servers with Nginx or Apache, DNS, HTTPS, systemd services, backups, and monitoring with Prometheus and Grafana dashboards.
Linux, Nginx, Docker, Prometheus, Grafana
I also build web applications in PHP, Laravel and Node.js. If your problem sits between code and infrastructure, I'd like to hear about it.
Talk to meThis is a typical delivery pipeline I build and run. Select a stage to see what happens there, with an example of the code behind it.
Every change starts as a commit to Git and goes through review. Pushing to the main branch triggers the pipeline, so nobody deploys from their own laptop.
$ git push origin main
# webhook triggers Jenkins job app/mainJenkins, GitHub Actions or GitLab CI checks out the commit and builds one artifact, tagged with the commit hash. That same artifact moves through every environment.
stage('Build') {
steps {
sh 'docker build -t app:${GIT_COMMIT} .'
}
}Automated tests run on every build. A failing test stops the pipeline before the change gets anywhere near users.
stage('Test') {
steps {
sh 'bundle exec rake test'
}
post {
always { junit 'reports/*.xml' }
}
}The environment is created or updated from Terraform or CloudFormation. The same code with different variables produces staging and production.
resource "aws_instance" "app" {
ami = var.ami_id
instance_type = "t3.small"
tags = {
Name = "app-${var.env}"
}
}
$ terraform plan -var env=staging
$ terraform apply -var env=stagingAnsible playbooks, Puppet manifests or Chef cookbooks set up packages, services and files on each server, which keeps every machine identical.
- name: Web servers
hosts: web
become: true
tasks:
- name: Install nginx
apt: { name: nginx, state: present }
- name: Keep nginx running
service: { name: nginx, state: started, enabled: true }ArgoCD syncs the approved version from Git to the cluster, so production always matches what's in the repository. Every release follows a plan and has a tested rollback path.
apiVersion: argoproj.io/v1alpha1
kind: Application
metadata:
name: app-production
spec:
source:
repoURL: https://git.example.com/app-deploy.git
path: overlays/production
syncPolicy:
automated: { prune: true, selfHeal: true }After release, Prometheus metrics, Grafana dashboards and CloudWatch alarms confirm the change is healthy. If they don't, the rollback plan runs.
# Error rate over the last 5 minutes (Prometheus)
sum(rate(http_requests_total{status=~"5.."}[5m]))
/
sum(rate(http_requests_total[5m]))If a task comes up twice, it becomes a script, and once it's stable it becomes part of the pipeline.
Infrastructure and configuration live in version control, so they can be reviewed, reproduced and rolled back.
Shipping small changes often is safer than shipping big ones rarely, and every release has a rollback plan.
Nine years of writing application code means I build pipelines around how developers actually work.
About twelve years in tech, going from fixing workstations to shipping web apps to running delivery pipelines in the cloud.
I automate the software delivery pipeline and run applications in the cloud.
I built and maintained websites and web applications from feature work through to production deploys.
I supported users, workstations and networks. Most of my troubleshooting habits come from this job.
Web Engineering Certificate, John Bryce College. High school diploma.
Arabic (native), Hebrew (fluent), English (professional).
The tools I use day to day, grouped by the part of the delivery chain they cover.
Work I do outside the day job.
Technical co-founder and CTO
ZVV is a real-estate platform that brings property listings, bookings, investments and an AI layer into one product. I own the technical side, which covers architecture, backend and API development, security and continuous integration.
A Node.js and Express chat service using the Groq SDK, with conversations stored in SQLite. It runs as a systemd service on a VPS behind Nginx with HTTPS, and has a health endpoint for monitoring.
My desktop setup lives in version control: Sway and Waybar, Kitty, Zsh, tmux and Neovim, with cron jobs for package backups so I can rebuild a machine quickly.
Quick answers to what recruiters and teams usually ask me first.
DevOps engineering, release and platform roles, and positions that combine infrastructure work with web development.
I'm based in Israel and open to relocation.
AWS. I use the AWS CLI and CloudFormation daily, along with EC2, S3, IAM, VPC and CloudWatch.
Yes. I worked as a web developer from 2014 to 2023 with PHP, Laravel, JavaScript and MySQL, and I build Node.js services in my own projects.
Jenkins for pipelines, AWS CloudFormation for infrastructure as code, Puppet and Chef for configuration management, and Ruby and Bash for automation scripts.
Arabic (native), Hebrew (fluent) and English (professional).
Email me at [email protected] or message me on LinkedIn.
I'm open to DevOps and platform roles, and to conversations about delivery pipelines, cloud infrastructure or web projects.
[email protected]