Growth Engineer Marketing Operations & Growth Systems

I build the systems behind growth.

I’ve spent 10+ years inside marketing, from product, app-first commerce and performance leadership to website operations and enterprise transformation. Today I work across marketing operations, designing and building the digital experiences, measurement, data products, integrations and AI workflows that help teams understand what happened, decide what matters and carry that context forward.

Follow the signal
L’Oréal istegelsin Getir Eczacıbaşı Roveir CEFA
Marketing judgment, measurement, internal products and governed AI converge into reusable operating context.

Systems are easier to build when you know the work up close.

Before I built tools, I worked inside the work itself: campaigns, app-first commerce, performance teams, websites, quality checks and reporting. Repetition made the weak points easy to see.

I began turning those weak points into shared definitions, clearer processes and useful automation. That path led through attribution, team leadership, CRM, commerce, data products and AI evaluation. I still work on growth. I build its operating layer with the day-to-day reality in view.

  1. Operate Campaigns, commerce, websites and daily QA
  2. Standardize Definitions, quality and ownership
  3. Build Tools, automation and integrations
  4. Learn Evidence returned to the work

The strategy and the implementation stay together.

I bring the commercial context from a decade inside marketing, then work through website operations, quality standards, measurement, integrations and product surfaces myself. The daily operation matters as much as the individual tools.

  1. 01

    Website and commerce operations

    Digital experience, acquisition, SEO, quality control and commerce work tied to the customer journey and the commercial decision behind it.

    Operate
  2. 02

    Measurement infrastructure

    Event contracts, source boundaries and governed identities that preserve the business context behind a number.

    Evidence
  3. 03

    Products and automation

    Purpose-built tools, APIs, integrations and operating views that remove repeated work without hiding the decision.

    Action
  4. 04

    Governed learning

    Source status, evaluation and feedback records that keep useful judgment available for the next cycle.

    Memory

Different businesses. One operating pattern.

App-first commerce made the commercial loop visible. Enterprise roles taught me how quality and shared standards travel across teams. The current systems work shows what happens when those lessons become tools, data foundations and operating products.

View the work index
Education network · 50+ locations

One operating discipline. Several working products.

A measurement problem expanded into connected work across data, CRM, parent experience, local visibility and commerce. The products are separate. The discipline behind them is shared.

Paid and organic Website inquiries CRM context Parent questions Location evidence
01 / Foundation Measurement + CRM

Preserve what the next system needs.

A first-party form contract, CRM-context handoff and governed network identity keep source, school and program meaning intact.

  • Event identity
  • Source boundaries
  • Canonical locations
  1. 02 / Operating view Decision support

    Network Performance View

    School, network, parent-insight and source-status signals in one invitation-only operating view.

    Network and school-level view
  2. 03 / AI evaluation Quality intelligence

    Conversation Diagnostics

    Intent, answer source, fallback and quality diagnostics for an observable chatbot improvement loop.

    Synthetic conversation records
  3. 04 / Local intelligence Search intelligence

    Local Presence Intelligence

    Four Ontario schools and 500 exact-place Maps checks, combined with review, page and demand context.

    Location-level search evidence
  4. 05 / Commerce product Shopify

    Commerce Rebuild

    A Shopify rebuild with custom storefront, school-location and bundle logic.

    Storefront and bundle logic

The work returns to the system.

Most projects start with a decision that has become hard to make. I trace the signal to its source, decide what can be trusted, build the smallest useful operating layer, and return the result to the next cycle.

  1. 01

    Observe

    Trace the decision back to its source.

  2. 02

    Model

    Establish what the evidence can support.

  3. 03

    Build

    Make the smallest useful operating layer.

  4. 04

    Measure

    Check what changed in the decision.

  5. 05

    Return

    Preserve the result for the next cycle.

I kept moving closer to the system.

The titles stay true to the roles. The throughline is the work: from product marketing and ad operations into measurement, team leadership, transformation, internal products and growth infrastructure.

  1. 01

    Mynet

    Istanbul

    Product Marketing Manager

  2. 02

    Hürriyet

    Istanbul

    Programmatic & Ad Operations Executive

  3. 03

    L’Oréal

    Istanbul

    Senior Precision Advertiser

    Digital operations · Four divisions
  4. 04

    istegelsin

    Istanbul

    Head of Performance Marketing

    App-first quick commerce · Team leadership
  5. 05

    Getir

    Western Europe

    Digital Marketing Manager

    Western Europe · Quick commerce
  6. 06

    Eczacıbaşı

    Istanbul

    Digital Transformation / Marketing Manager

  7. 07

    Roveir

    Vancouver

    Growth Lead

  8. 08

    CEFA Early Learning

    Vancouver

    Paid Media and SEO Specialist

    Multi-location systems work

Point of view

I’m most useful when a business has more data than shared understanding. I make the source, owner and assumption visible, then build the smallest product that helps people decide. Automation handles the repeated work. The judgment stays where people can inspect it.

Independent concept

Creative Experiment Memory

A working model for keeping a hypothesis, brief, asset, audience, result and next decision in one record. Every example is synthetic.

Open the experiment