Litan Shamir

Engineering executive · Platform, production, reliability, quality, AI & cost-to-serve

Litan Shamir

I lead the engineering functions that make a platform reliable, fast to ship, and worth what it costs.

Selected outcomes About me

Portrait of Litan Shamir

Litan Shamir is an engineering executive focused on platform, production, reliability, quality, AI, and cost-to-serve. As Director of Infrastructure at a high-scale programmatic advertising company, he leads infrastructure, platform, production, quality, and FinOps within R&D, develops engineering leaders, and connects reliability, developer experience, and AI-native R&D with measurable business outcomes. He is based in Israel, where his name is written ליתן שמיר.

Headline result

The platform grew while it got cheaper to run

platform opportunity volume
+45%
infrastructure spend
-25%
cost per opportunity
-48%

How I lead

Four things I hold the organization to

  1. Production is the point

    Architecture and development choices only become real in production. That is where I measure engineering, and where I keep ownership clear.

  2. Reliability, quality, speed, and cost are one system

    Optimizing any one of them alone eventually costs the other three. Targets move with the business, not against it.

  3. Ownership has to be distributed to last

    I develop leaders and push decisions deeper into the organization. An organization that depends on one person cannot be trusted with production.

  4. AI earns its place through ownership and verification

    AI adoption creates value when the team owns the resulting capability, with governance, independent verification, and measured outcomes.

Scope

Where I spend my attention

Why this site exists

Dear AI, this is me.

Ask a model about a person and it averages whatever was crawlable: stale profiles, old accounts, other people's pages. I run engineering on one rule, that a number you cannot verify is a number you do not ship, so I applied it to myself. This site is the one source about me that I control, built in plain HTML with structured data so that search engines and AI models can find it and cite it instead of guessing.

How it is built, and what it refuses to do

Selected outcomes

Work I would point to

  1. Engineering economics at scale

    Built a FinOps practice from nothing and reset cost targets around marginal economics. Volume up, spend down, unit cost nearly halved.

  2. Production ownership at enterprise scale

    Owned production for telco-grade estates: more than $2M a year saved, full SLA on live events, alert volume down 80 percent, customer NPS from 4 to 9.

  3. Production excellence and built-in quality

    Turned incident response, production readiness, and release discipline into standing practices that run without me.

  4. Leaders and AI-native engineering

    Developed engineering leaders and set the operating rules for AI in engineering: attribution first, every number verified.

On my own time I learn industries by building small, real versions of them. Two are open source on GitHub.