Machine Learning Engineer
San Francisco, CA
ViacomCBS Streaming (formerly CBS Interactive) is a division of ViacomCBS that encompasses both free, paid, and premium streaming services including Pluto TV, CBS All Access, CBS Sports Digital, and CBS News Digital.
Interest in applying machine learning e.g, supervised methods like logistic regression, SVMs, random forests, deep learning reinforcement learning, to existing products and services to improve their effectiveness and the overall user experience across our digital properties.
Leverage datasets like traffic/video/ consumption data from Adobe Analytics clickStream, purchase/ subscription data, audience data from syndicated audience measurement services (comScore/ Nielsen), and Ad revenue data from Doubleclick, to model, analyze and predict user behavior.
Utilize various statistical measures (mean, median, variance, etc.), distributions (uniform, normal, binomial, Poisson, etc.) and analysis methods (ANOVA, hypothesis testing, etc.) to build and validate models from observed data.
• Build and implement production machine learning systems
• Maintain the health of machine learning systems, including speed, reliability, and performance
• Develop internal machine learning frameworks and abstractions to facilitate common tasks such as training/testing, feature use/reuse/creation/storage, and deployment. - These abstractions are used by both machine learning engineers and data scientists/BI.
• Design machine learning systems
• Leverage impactful and innovative technology based on machine learning to improve existing products like recommendation playlists, and create new products like implicit preferences-based related-story widgets, user-optimized content on the homepage, and smart onsite messaging tools to present users with the perfect offer at the perfect time
• Use your mathematical, statistical and programmatic knowledge to search, analyze, and model in order to forecast the traffic and consumption of our digital properties and the sale of our subscription products
• Read white papers, synthesize information, apply theoretical ML concepts to real-world business problems across our digital properties
• Document, train, and mentor members of the Data and BI teams on ML concepts and principles, and how they can be applied to solve day-to-day business needs
• Work on initiatives that improve the quality of our data as well as making the data more effective, with other members of engineering, BI teams, and business units to implement changes.
• Build a user-level data management platform that serves as the central data hub allowing machine learning software components to ''plug-in'' and segment users, predict future purchase behavior, identify users at risk of churning, and personalize their online experiences.
• Use look-alike modeling derived from Adobe Analytics clickStream-sourced behavioral data to build larger audience segments from smaller segments in order to create targeted ''reach'' for advertisers
• Work on device/user graph that accounts for multi-platform access, and pan-visit user behavior
• Incorporate ''anomaly detection'' across KPIs, metrics, and datasets to proactively find data issues, data load problems, and failures with minimal human supervision
What you bring to the team:
You Have -
• MA/MS in Statistics/Data Science/Computer Science or related disciplines with specialization in data mining or machine learning techniques
• Knowledge of both supervised and unsupervised machine learning techniques
• Have full stack experience in data collection, aggregation, analysis, visualization, productionalization, and monitoring of data science products
• Proficiency in big data modeling work: BigQuery, HBase
• The ability to write robust code in Python and associated machine learning packages is a must.
• Experience using Jupyter Notebooks
• Communicate concisely and persuasively with engineers and product managers
• Strong detail-orientation with a penchant for data accuracy and good grammar
• Familiar with version control systems (Git, BitBucket)
• Must successfully pass a background check
You might also have -
• Familiarity in Hadoop pipelines using Spark, Kafka.
• Familiarity with Graph Databases (Neo4J,
• Experience using project management tools like those from Atlassian (JIRA, Confluence)
• Experience using Google Cloud Platform (BigQuery, ML Engine, and APIs)
• Can Wrangle data using: SQL, Pandas, MongoDB
• Background in NLP or text mining techniques is a plus
• Background in deep learning and Tensorflow is a plus
FUNCTION: Data and Research
Jobcode: Reference SBJ-gkkz50-3-236-122-9-42 in your application.
CBS Interactive is the premier online content network for information and online operations of ViacomCBS as well as some of the top native digital brands in the entertainment industry.