PyMC-Marketing MMM Engineer

About the Project

We are an early-stage healthcare measurement startup building a privacy-safe marketing attribution and measurement product for healthcare organizations.

The product is live, and we have a paying client lined up. We are looking for an experienced PyMC-Marketing / PyMC MMM engineer to build and tune our first production marketing mix model using real client data.

The selected contractor will work with de-identified aggregate marketing data, such as spend, conversions, and aggregate call counts. This role does not involve handling PHI. 

This is a fixed-scope, 4-week project with potential for follow-on project work if the initial build is successful.

**Please note - Engagement Details

• Duration: 4 weeks
• Time commitment: approximately 20 hours per week
• Compensation: fixed-price project, $7K–$9K range
• Location: remote / global
• Follow-on work: potential for additional fixed-scope projects or retainer-style phases after the initial build

Deliverables

By the end of the 4-week project, the selected contractor should deliver:

• Working PyMC-Marketing / PyMC MMM pipeline
• Initial model fit using real client marketing and call-tracking data
• Calibrated model assumptions and priors
• Outputs for channel ROI, marginal return, and budget recommendations
• Lightweight Streamlit dashboard showing model results
• Validation approach, including holdout, calibration, incrementality, or other relevant methods where applicable
• Documentation covering inputs, methodology, assumptions, limitations, and recommended next steps

Project Scope

You will build the first production MMM pipeline for a healthcare marketing use case.

The project will include:

• Standing up a PyMC-Marketing pipeline in AWS SageMaker using a provided container
• Ingesting 18–24 months of Google Ads, Meta Ads, and call-tracking data
• Fitting, tuning, and calibrating the first model
• Applying healthcare-appropriate priors and assumptions
• Producing model outputs around channel ROI, marginal return, response curves, and budget recommendations
• Building a lightweight Streamlit dashboard to display key model outputs
• Clearly documenting what the model can and cannot tell the business
• Documenting model assumptions, limitations, inputs, outputs, and handoff notes

Required Experience

This role requires hands-on production experience with PyMC-Marketing or PyMC-based MMM / time-series modeling.

You should have experience with:

• Building and shipping a production MMM or Bayesian time-series model using PyMC-Marketing or PyMC
• Working with real client or business marketing data
• Marketing mix modeling concepts including adstock, saturation, priors, posterior uncertainty, ROI, marginal return, and budget optimization
• Validating models against real-world outcomes, holdouts, calibration data, incrementality tests, or other ground-truth signals
• Python-based modeling workflows
• Translating model outputs, uncertainty, and limitations into business-facing recommendations
• Working independently in a fast-moving startup environment

Nice-to-Haves

• Healthcare, regulated industry, or privacy-safe marketing measurement experience
• Experience with AWS SageMaker or similar cloud-based model environments
• Streamlit dashboard development
• Experience with Google Ads, Meta Ads, call-tracking, or lead-generation data
• Prior work with healthcare marketing, patient acquisition, or call-based conversion funnels