I'm a DevOps engineer based in Tel-Aviv, currently at Inuitive (Ra'anana), building CI/CD pipelines and developer tooling for teams that ship computer vision chips.
My work sits at the intersection of build systems, cloud infrastructure, and automation — from migrating legacy SCM to Git and rebuilding C/C++ packaging with Conan, to running Jenkins pipelines and designing DevEx tools that get out of engineers' way.
Before DevOps I spent five years as a Company Commander in the IDF — managing people, logistics, and 2M ILS budgets. That experience shapes how I think about systems: reliability is a people problem before it's a technical one.
Led the full SCM migration from SVN to Git across the org — preserving history, handling large binary assets with Git LFS, and splitting monolithic repositories into subdirectories. Wrote the internal migration guide that got the whole team across with minimal disruption.
Migrated Visual Studio solutions and Ninja builds to CMake, and introduced Conan as the C/C++ package manager. The result: reproducible builds, consistent dependency versions across machines, and a packaging workflow that scales.
Built and maintained the Continuous Integration pipeline used in production at Inuitive — from source commit to artifact. Integrated JFrog Artifactory as the artifact repository, streamlining the build, test, and deployment flow for hardware-adjacent software.
Designed and shipped Python-based developer experience tools with lightweight UIs and REST APIs, cutting manual steps for the engineering team. Also created and deployed a Debian PPA repository on AWS to distribute internal packages cleanly across environments.
Velocity points, commit counts, PR throughput — I tracked all of it. Then I noticed the team with the best metrics was the most miserable, and the best team had the worst numbers. Here's what I measure now.
Good documentation is a failure mode. What if the system's behavior was the documentation? On making infrastructure legible without a wiki.
Every few years someone declares Unix is dead. Every time, they're wrong. Why one idea from 1978 still shapes how I design everything.
The best systems I've built had strict limitations. I've started deliberately introducing constraints on projects that don't have them.
I came to DevOps through the army. Five years as a Company Commander in the IDF — managing people, coordinating logistics, running budgets — taught me that the hardest part of any system is the human layer. I brought that lens into engineering when I joined Inuitive in 2022.
At Inuitive I work on the pipeline that takes code from commit to silicon. Build systems, CI/CD, packaging, developer tooling — the infrastructure that lets chip engineers ship without friction. In parallel I'm completing a BSc in Computer Science at the Open University of Israel.
My tech comfort zone: Jenkins, Python, Bash, Docker, Kubernetes, Terraform, and AWS. I hold the AWS Certified Cloud Practitioner (CLF-C02) and I'm comfortable across Linux, Windows, and macOS. I believe in automation that explains itself and tooling that gets out of the way.