Fractional Data Scientist – Marketing Mix Modeling

About the role

We are building a privacy‑first healthcare marketing attribution platform and need a Fractional Data Scientist to design the statistical engine behind our Media Mix Modeling (MMM) product. You will create a Bayesian modeling framework that turns aggregate, privacy‑safe marketing data into actionable insights for budget allocation.

Key responsibilities

• Design and implement an automated Bayesian MMM framework, evaluating tools such as Google Meridian, PyMC, NumPyro or similar.
• Develop methods for calibrating Bayesian priors using historical data and micro‑holdout or geo‑testing variance.
• Incorporate ad‑stock effects, decay, and non‑linear saturation curves to identify optimal spending thresholds.
• Define input schemas, output structures, assumptions and limitations of the model.
• Build privacy‑safe attribution logic that works with server‑side tracking and aggregate data.
• Design budget‑optimization rules to flag diminishing returns and recommend reallocations.
• Generate probabilistic reports with median estimates and confidence intervals.
• Package model code and documentation for seamless hand‑off to the engineering and dashboard teams.

Required profile

• Proven experience in statistical or Bayesian modeling for marketing mix analysis.
• Ability to work independently and collaborate with founders, developers and designers.
• Strong understanding of privacy‑first data aggregation and server‑side tracking concepts.

Required skills

• Python programming
• Probabilistic programming libraries (PyMC, NumPyro, Google Meridian)
• Bayesian statistics and prior calibration
• Ad‑stock, decay and saturation modeling
• AWS serverless data pipelines
• Privacy‑safe attribution methodologies