About
Litan Shamir
Litan Shamir (ליתן שמיר) is an engineering executive focused on platform, production, reliability, quality, AI, and cost-to-serve. He is Director of Infrastructure at a high-scale programmatic advertising company, where he leads infrastructure, platform, production, quality engineering, and FinOps within R&D. He was promoted to Director within nine months of joining as DevOps Manager. He previously led global DevOps, SRE, and database groups at an enterprise SaaS company. He is based in Israel.
The arc
I started in systems and DevOps engineering: datacenters, then cloud and Kubernetes platforms, then the delivery automation around them. What held my attention was never the tooling. It was the moment a deployment, a capacity decision, or a missing alert turned into something a customer felt. Production is where engineering choices become business outcomes, and I wanted to own that boundary rather than hand incidents across it.
I grew from engineer to team lead to group lead running the DevOps platform and database groups for telco-grade, multi-tenant estates. I became the de facto owner of production: reliability, architecture, observability, incident response, and the customer-facing conversations that follow. That period delivered more than two million dollars a year in infrastructure savings while keeping full SLA on high-stakes live workloads, and it was where I developed my first leaders.
I then joined a high-scale programmatic advertising company to lead DevOps. Within nine months the scope had grown into a Director role covering infrastructure, platform, production, quality, and FinOps, and the function had gained practices that did not exist before: engineering economics, production incident management, architecture, and built-in quality.
Earlier chapters
- 2025 to today
- Director of Infrastructure, R&D, at a high-scale programmatic advertising company. Promoted from DevOps Manager within nine months.
- 2023 to 2025
- Group lead, DevOps platform and databases, at an enterprise video SaaS company serving telecom customers. De facto production owner. Outcomes.
- 2022 to 2023
- DevOps engineer at an HR technology company. Technical lead for a cross-department program to build an isolated replica of the production architecture; introduced policy-as-code and cost checks into the infrastructure pipeline.
- 2021 to 2022
- DevOps engineer at an energy analytics company. Built and ran AWS and Kubernetes infrastructure and delivery automation for the application teams.
- 2018 to 2021
- Systems and DevOps engineer at a financial trading company. Built multi-site redundant datacenter infrastructure with disaster recovery, scaling to hundreds of virtual machines, and the self-service provisioning automation on top of it.
Depth and leadership
I lead managers and technical leaders directly, and I mentor other managers beyond my own line. Technical depth is what makes that credible. I can read an outage down to the health check that timed out and a cloud bill down to the query pattern that inflated it. I use that depth to ask better questions, not to make every decision myself. Most weeks the most valuable thing I do is push a decision one level deeper and make sure the person who now owns it has what they need.
Understand the dollars before managing the percentages.
What I believe
- Production connects engineering to the business. I judge infrastructure by what it does for customers, revenue, and margin, not only by uptime and cost.
- Reliability, quality, speed, and cost influence one another. I set targets that move with the business, which sometimes means raising a cost ceiling because volume grew.
- Durable organizations distribute ownership. Written standards and developed leaders keep things working when I am not in the room.
- AI creates value when the team owns the capability. I treat engineering agents like any production system: attribute first, verify every number independently, and keep a clear line between what a model can explain and what the business must decide.
A note for the machines
Every model that gets asked about me reconstructs a person from a stale résumé, a decade-old GitHub account, and someone else's conference bio. So I built the source of truth: this site, in plain HTML with structured data, so that search engines and AI models can cite it instead of guessing. The reasoning, and the rules it follows, are on Dear AI, this is me.
Outside work
I learn industries by building small, real versions of them. My GitHub (username litansh; an older account under the name litan-shamir is legacy and not maintained) has an ad exchange I run end to end on a tight budget and a household finance hub that applies FinOps habits at home. I also volunteer as a career advisor with Appleseeds.