Company Description
Blend is building a scalable Media Mix Optimization (MMO) solution designed to help clients maximize the impact of their marketing investments. We are seeking a Data Scientist with strong expertise in media mix modeling, statistical modeling, and interactive application development to join our advanced analytics team. This role goes beyond model building — you will design, implement, and productionize end-to-end solutions that integrate statistical rigor with business impact.
The ideal candidate will have deep knowledge of marketing analytics, advanced Python skills, and hands-on experience with Streamlit or similar frameworks for interactive data applications. You will be central in creating robust pipelines, experimentation frameworks, and client-facing tools that directly inform media allocation decisions.
Job Description
Project Overview: Media Mix Optimization (MMO)
Our MMO platform is an in-house initiative designed to empower clients with data-driven decision-making in marketing strategy. By applying Bayesian and frequentist approaches to media mix modeling, we are able to quantify channel-level ROI, measure incrementality, and simulate outcomes under varying spend scenarios.
Key components of the project include:
Data Integration: Combining client first-party, third-party, and campaign-level data across digital, offline, and emerging channels into a unified modeling framework.
Model Development: Building and validating media mix models (MMM) using advanced statistical and machine learning techniques such as hierarchical Bayesian regression, regularized regression (Ridge/Lasso), and time-series modeling.
Scenario Simulation: Enabling stakeholders to forecast outcomes under different budget allocations through simulation and optimization algorithms.
Deployment & Visualization: Using Streamlit to build interactive, client-facing dashboards for model exploration, scenario planning, and actionable recommendation delivery.
Scalability: Engineering the system to support multiple clients across industries with varying data volumes, refresh cycles, and modeling complexities.
Responsibilities
Develop, validate, and maintain media mix models to evaluate cross-channel marketing effectiveness and return on investment.
Engineer and optimize end-to-end data pipelines for ingesting, cleaning, and structuring large, heterogeneous datasets from multiple marketing and business sources.
Design, build, and deploy Streamlit-based interactive dashboards and applications for scenario testing, optimization, and reporting.
Conduct exploratory data analysis (EDA) and advanced feature engineering to identify drivers of performance.
Apply Bayesian methods, regularization, and time-series analysis to improve model accuracy, stability, and interpretability.
Implement optimization and scenario-planning algorithms to recommend budget allocation strategies that maximize business outcomes.
Collaborate closely with product, engineering, and client teams to align technical solutions with business objectives.
Present insights and recommendations to senior stakeholders in both technical and non- technical language.
Stay current with emerging tools, techniques, and best practices in media mix modeling, causal inference, and marketing science.
Qualifications
Bachelor’s or Master’s degree in Data Science, Statistics, Computer Science, Applied Mathematics, or related field.
Proven hands-on experience in media mix modeling, marketing analytics, or econometrics.
Strong proficiency in Python and key data science libraries (pandas, NumPy, scikit-learn, statsmodels, PyMC or similar Bayesian frameworks).
Experience with Streamlit or equivalent frameworks (Dash, Shiny) for building data- driven applications.
Proficiency in SQL for querying, joining, and optimizing large-scale datasets.
Solid foundation in statistical modeling, regression techniques, and machine learning.
Strong problem-solving skills with the ability to structure ambiguous business problems
into data-driven solutions.
Excellent verbal and written communication skills to translate technical outputs into
business decisions.
Preferred Qualifications
Experience with Bayesian hierarchical models, time-series decomposition, and marketing attribution approaches.
Familiarity with cloud-based platforms (AWS, GCP, Azure) for data processing, model training, and deployment.
Experience with data visualization tools beyond Streamlit (Tableau, Power BI, D3.js, Plotly).
Exposure to big data ecosystems (Spark, Hadoop) for large-scale data processing.
Knowledge of causal inference techniques (propensity scoring, uplift modeling, geo-
experiments).