Remote | $100/hour | W-2 or 1099 | Contract-to-Hire

THE OPPORTUNITY

Our client is a growing performance marketing and media agency managing more than $10 million in annual media for 30+ clients across CPG, healthcare, DTC, and B2B.

The agency is expanding a proprietary analytics platform that helps clients understand what is truly driving performance and where to invest next. They are looking for someone to build and own the Bayesian marketing measurement methodology behind it.

This is not a role for someone who simply operates an existing MMM tool or manages an outside vendor. You will create the framework from the ground up, work directly with the CEO and senior client team, and have a meaningful voice in how the capability develops.

The environment is small, fast-moving, and highly visible. Your work will directly influence client media decisions, with a path to a senior full-time role as the platform and client base grow.

THE ROLE

You will architect, build, and operate a proprietary marketing measurement framework based on Bayesian Seemingly Unrelated Regressions, hierarchical estimation, and adaptive media structuring.

Responsibilities include:

• Build a Bayesian SUR framework using PyMC, Stan, NumPyro, or an equivalent platform.
• Develop hierarchical models and partial-pooling structures across channels, campaigns, creatives, audiences, and placements.
• Implement adstock, saturation, carryover, and synergy transformations.
• Develop multi-KPI analysis across awareness, engagement, intent, and conversion.
• Establish rigorous model validation, including posterior predictive checks, rolling holdouts, experimental calibration, and coefficient-stability diagnostics.
• Produce media contribution estimates, saturation curves, marginal efficiency analysis, and budget optimization scenarios.
• Translate complex findings into clear recommendations for clients and media teams.
• Maintain and evolve the framework as new clients, channels, and data sources are added.

WHAT WE ARE LOOKING FOR

Required:

• Graduate-level training in econometrics, statistics, economics, or another quantitative field.
• Hands-on experience with Seemingly Unrelated Regression systems.
• Proficiency with Bayesian probabilistic programming using PyMC, Stan, NumPyro, or an equivalent platform.
• Strong knowledge of hierarchical modeling, partial pooling, and correlated time-series data.
• Professional experience with marketing mix modeling, media measurement, attribution, or marketing science.
• Strong Python skills and the ability to explain statistical findings to non-technical stakeholders.

Preferred:

• PhD in economics, statistics, econometrics, or a related field.
• Experience with a marketing science firm, econometric consultancy, media agency, or advanced MMM practice.
• Experience with geo-lift, conversion-lift, or holdout experiments.
• Familiarity with advertising platforms and APIs such as Meta, Google, DV360, or The Trade Desk.

WHO WILL THRIVE HERE

You want to build something rather than maintain someone else’s system. You care about the statistical architecture behind marketing measurement, have opinions about aggregation bias, and are comfortable owning important modeling decisions.

This is a fully remote opportunity paying $100 per hour. It can be structured as either W-2 or 1099 and will begin as a fractional or contract engagement, with the potential to become a senior full-time role.