Skip to content
Amir Tavasoli

Executive Proof

Selected outcomes, packaged for hiring teams.

These case studies focus on business context, operating decisions, and measurable outcomes. They are written to show how I lead data science programs, not just how I build models.


Discuss an Opportunity

Sobeys Inc. | Jan 20, 2026

Leading agentic AI for offer creation

Led context-aware offer generation and personalization to reduce manual campaign planning effort and accelerate launch speed.

Role

Director of Data Science

Problem

Campaign offer planning relied on manual, repetitive workflows that slowed launches and made personalization harder to scale consistently.

Agentic AIPersonalization strategyMarketing operationsAI product leadershipCross-functional execution

Sobeys Inc. | Nov 12, 2025

Leading agentic AI plus MMM advisory decisioning

Led an AI decision layer over MMM outputs that translated model signals into budget and channel allocation guidance for marketing leadership.

Role

Director of Data Science

Problem

MMM outputs were analytically strong but difficult to convert into timely, actionable budget and channel decisions for senior stakeholders.

Agentic AIMarketing Mix ModelingDecision systemsExecutive communicationAI strategy

Sobeys Inc. | Jan 15, 2025

Building a loyalty decision engine at retail scale

Reframed personalization from disconnected model outputs into an executive-level decisioning system for loyalty and marketing teams.

Role

Director of Data Science

Problem

High-value personalization work was fragmented across channels, creating inconsistent execution and limiting measurable business lift.

AI strategyPersonalizationExecutive stakeholder managementOperating model designDecision systems

Sobeys Inc. | Jan 18, 2024

Standing up experimentation as a growth system

Built the measurement and experimentation discipline needed to evaluate personalization investments with more confidence.

Role

Data Science Manager, Personalization

Problem

The organization needed a more credible way to decide which ML and personalization initiatives deserved continued investment.

Experimentation strategyMeasurement designPortfolio prioritizationExecutive communication

Sobeys Inc. | Sep 12, 2023

Reducing incident risk in personalization delivery

Introduced reliability practices that turned a fragile personalization workflow into a steadier production capability.

Role

Data Science Manager, Personalization

Problem

Frequent operational interruptions made it harder for the business to trust personalization outputs and plan around ML-enabled delivery.

ML reliabilityProduction operationsCross-functional process designTeam leadership