Hi, I'm Shobeir.

What I do now

I lead data science at CVS Media Exchange (CMX), building agentic marketing and retail media at the scale of a Fortune 6. My team designs AI agents that run campaigns end-to-end — analyzing real-time signals, adapting to consumer behavior, and pushing measurement, targeting, and audience automation past where traditional martech stops. The launch of CorIQ™, our AI-powered media platform built on ExtraCare® data, is the most recent product of that 18-month vision.

Earlier in my CVS run I built the personalization and recommendation engine for CVS.com, serving over 2.5 billion visitors, and helped install a real culture of experimentation across the org — making "test and learn" the default, not a side project.

How I got here

A decade-plus in data science across very different shapes of the problem:

• CVS Health — Executive Director, Data Science (CMX). Previously Director of Data Science & Personalization, and Senior Manager at Aetna/CVS.
• Gopuff — Principal Data Scientist, building the experimentation program (20× more experiments shipped, A/B + switchback platforms).
• Verra Mobility — Data Scientist on computer-vision and ops problems.
• Arizona State University — Instructor.
• Cheeky Scientist — Lead Program Developer for the Data Science Syndicate, mentoring PhDs into industry.
• Harvard Business Review — Advisory Council member since 2021.
• AMIA — Reviewer, American Medical Informatics Association.

Before data science

I came in through the side door. PhD in Physical Chemistry from Arizona State University (2015) — dissertation on molecular models for conductance in junctions and electrochemical electron transfer; M.Sc. in Physical Chemistry from Sharif University of Technology. Research spanned nickel catalysts for hydrogen production (with Dr. Jones' group, published in JACS) and non-equilibrium Green's-function transport in molecular wires. On the side I built AstroPaint, an open-source Python package for painting astrophysical signals onto sky maps. The quantum/astro detour shaped how I think about uncertainty, models, and what it means for a result to be "real."

What I'm into

Personalization, healthcare, experimentation, on-device AI, and small tools that turn a one-evening idea into something I can actually use. ai, may I? is the place I keep those experiments — illustrated public-domain classics retold with AI, a card-stitcher for scanned greetings, an on-device photo organizer (Glimp), and whatever's next.

Outside of all that, I draw and paint. @ink.visions is the visual journal.

Get in touch

If something here resonates — collaboration, a question about the work, a job in personalization or applied ML — shobmaz@gmail.com is the fastest way to reach me. LinkedIn works too.