What You'll Do:
- Own end-to-end delivery of significant data science projects — from problem scoping and approach design through to production deployment
- Make sound, independently-reasoned decisions on methodology, model selection, and evaluation; document them clearly in technical solution documents covering problem statement, approach, metrics, and timeline
- Lead solution design for your own initiatives; break down complex epics into well-scoped user stories with clear acceptance criteria, adopting DataOps and MLOps best practices throughout — experiment tracking, pipeline orchestration, model monitoring, and reproducibility
- Build production-quality Python and PySpark code on Databricks — well-tested, documented, and reusable — and implement advanced ML and AI-powered workflows including entity resolution, probabilistic record linkage, embedding-based matching, semantic similarity, and LLM-augmented pipelines
- Develop and maintain reusable tools, libraries, and documentation that improve team efficiency and technical standards; conduct code reviews with constructive, specific feedback that raises the bar
- Mentor junior data scientists on technical execution, code quality, and career development; lead internal talks or workshops on ML topics
- Collaborate cross-functionally with product, engineering, and operations — translate business requirements into technical specifications, partner with data engineering on scalable pipeline design, and participate in cross-functional design reviews and working groups
Who You Are:
- Bachelor's degree required in Statistics, Data Science, Computer Science, Mathematics or a related quantitative field; Master's strongly preferred
- 3–5 years of hands-on data science experience with demonstrated ability to own and deliver complex, multi-sprint projects independently
- Advanced Python with production-quality code, testing, and documentation; strong SQL and PySpark for billion-row datasets
- Databricks workflows, Delta Lake, and job orchestration; working knowledge of cloud platforms (AWS or GCP)
- Solid command of core ML — regression, classification, clustering, model evaluation, and experimental design — applied to complex, high-volume data
- Proficiency with MLOps practices: experiment tracking, pipeline orchestration (Airflow), and reproducible model deployment
- Exposure to modern AI methodologies: RAG systems, LLM-augmented models, vector databases, and semantic search
- Strong communicator — able to translate technical work into clear documentation, user stories, and cross-functional conversations
- Demonstrated ability to mentor junior data scientists and contribute to team standards
Preferred skills:
- Hands-on experience with knowledge graph construction, entity resolution, or semantic data modeling (RDF, OWL, SPARQL, or equivalent graph frameworks)
- Familiarity with probabilistic record linkage, identity graph approaches, or embedding-based entity matching at scale
- Experience with causal inference methods (A/B testing, synthetic control, uplift modeling)
- Experience with deduplication, enrichment, or web-to-TV linkage problems
- Background in media, ad tech, or measurement — TV viewership (ACR/STB data), digital audience modeling, cross-platform measurement (linear + CTV/OTT), or identity resolution in privacy-constrained environments
- Familiarity with the measurement and identity vendor landscape (Nielsen, Comscore, LiveRamp, The Trade Desk
What We Do
Television remains a vibrant cultural influence and an essential source of entertainment and information worldwide. Tremendous growth in content choices, and viewing platforms that allow us to watch anything, anytime, on any screen, has actually made it harder for viewers to discover and keep up with all the great programming available. It’s also more competitive for content providers to keep your attention, and for marketers to make strong, measurable connections with their target consumers. Technology that improves the viewing experience, enables content discovery, and addresses audience fragmentation across screens will strengthen television’s business model and relevance to consumers. Data is at the center of any solution to make TV better. Samba TV's technology is built into Smart TVs and easily maps to smart phones and tablets. By recognizing what's on screen, Samba TV learns what viewers like and using machine learning algorithms, enables discovery of shows and actors in a whole new way. Likewise, our data and measurement products are transforming the way stakeholders across the media landscape are thinking about their business. Given the dramatic growth in streaming services, connected devices, time-shifting, and multi-screen viewership, our data products solve real problems and create a meaningful competitive advantage for our clients.
