Data Scientist 3 - Comcast Advertising

Comcast Advertising is driving the TV advertising industry forward, from delivering ads to linear and digital audiences to pioneering the tech that makes it possible. We help brands connect with their audiences on every screen using advanced data, technology, and premium video content. Our media sales division helps local, regional, and national brands reach potential customers through multiscreen TV advertising. Our ad tech division FreeWheel provides comprehensive adtech that makes it easier to buy and sell premium video advertising across all screens, data types, and sales channels.

This job involves extracting insights from complex, high-dimensional data using advanced data science techniques. It requires developing predictive models and analytical solutions to boost business performance. The role is key for driving data-informed business strategies and decisions.

Responsibilities: Collaborating with cross-functional teams to define data-driven questions and conduct experiments for product and service enhancements Engineering and refining algorithms, software, and automated processes for robust data integration and cleansing from multiple sources Utilizing statistical rigor and advanced data science methods to analyze large datasets, deriving insights through predictive models and machine learning Developing statistical and mathematical solutions to complex business problems, contributing to broader initiatives with minimal supervision Architecting data mining models and protocols to discern trends within extensive, intricate datasets, enhancing customer and product insights Deploying predictive analytics based on historical data to anticipate customer behavior and support strategic business decisions Synthesizing forecasts and strategic recommendations by applying data science to business data for impactful project support Advancing business decision-making by researching and applying novel data science principles and emerging analytical techniques Consistent exercise of independent judgment and discretion in matters of significance. Regular, consistent and punctual attendance. Must be able to work nights and weekends, variable schedule(s) as necessary. Other duties and responsibilities as assigned.

Experience & Impact: 5-8 years of broad marketing experience both in business areas (e.g., Sales, Yield, Product Development) as well as multiple years in areas that apply machine learning to real-world business problems, with a track record of leading AI-ML model development and taking ideas from concept to production. Core Machine Learning: Deep mastery of AI-ML techniques like tree-based analysis (XGBoost, Random Forest), Regression, Classification, Natural Language Processing, and/or Time-series forecasting. Standard evaluation metrics Knowledge of how to evaluate models and how to work with business to set service levels (RMSE, MAPE, MAE, R-squared, etc.) Programming & SQL: Strong Python proficiency (pandas, NumPy, scikit-learn) and advanced SQL (window functions, query optimization, complex joins). Experimental Design & A/B Testing: Practical knowledge of designing, executing, and analyzing randomized control trials, power analysis, and hypothesis testing. Communication & Stakeholder Management: Proven ability to translate technical modeling decisions into actionable business insights for non-technical stakeholders. Bachelor’s degree: Mathematics, Statistics, Computer Science, Data Science, Econometrics, Psychometrics, or related mathematical field. Graduate degree preferred.

Production MLOps: Experience with model deployment frameworks (Docker, MLFlow, GIT, AWS CLI). Advanced Architecture: Hands-on experience with deep learning frameworks (PyTorch and/or TensorFlow) Generative AI – Familiarity with LLM fine-tuning and RAG pipelines.  Familiarity with Agentic. Distributed Stack: Familiarity with distributed computing platforms like PySpark. Domain