Senior Data Scientist

Samba connects the advertising industry. Today, a trillion dollars of media flows through platforms that don’t talk to each other, and each one measures its own results. We built the independent alternative: consented first-party data and AI that give the world’s biggest advertisers, publishers, and platforms one clear view across all of it.

The next era of advertising will run on AI agents, and those agents will need data they can trust. We’re building that foundation, and the agents that act on it. Join us and help build what the industry runs on.

You help shape Samba’s data science function for the agentic era of advertising, defining modeling methodology, building production ML systems, and applying modern AI and agentic capabilities across our data products. You are an active collaborator with engineering, product, and partner teams, and an advocate for rigorous, defensible modeling practice.

• Own Delivery: Own end-to-end delivery of data science projects, from problem scoping through production deployment, building production-quality Python and PySpark on Databricks that is well-tested, documented, and reusable.
• Define Methodology: Define and ship the modeling methodology behind Samba’s data products, including model selection, evaluation frameworks, and reproducibility standards, applying core ML and statistics to billion-row, real-world data.
• Partner on Pipelines: Partner with Data Engineering to define data requirements, validate pipelines, and ensure model inputs are reliable and production-ready.
• Build AI Systems: Build and operate advanced AI systems using retrieval-augmented generation (RAG), large language model (LLM) augmented modeling, and Graph Neural Networks that improve as signals evolve and support agentic decision-making.
• Integrate Agentic Workflows: Integrate LLMs and agentic workflows into production ML pipelines where they extend modeling capability and unlock new product surfaces.
• Design and Mentor: Drive technical design for modeling components, translate business requirements into modeling solutions with product, engineering, and go-to-market partners, and mentor data scientists through feedback, pairing, and design review.
• Run MLOps: Establish and operate MLOps practices (experiment tracking, Airflow orchestration, model monitoring, retraining workflows) and apply privacy-compliant data handling, including GDPR, CCPA, and Samba’s data governance policies.

• 8+ years of hands-on data science experience with a Bachelor’s degree in Statistics, Data Science, Computer Science, Mathematics, or a related quantitative field (6+ with a Master’s, 3+ with a PhD), or equivalent practical experience.
• A demonstrated ability to own and deliver complex, multi-sprint data science projects from problem scoping through production deployment, and a track record of building methodology, not just applying it.
• Solid command of core ML and statistics, including neural networks, regression, classification, clustering, model evaluation, experimental design, and causal inference, applied to billion-row datasets.
• Advanced Python with production-quality, tested code, strong SQL and PySpark, and hands-on production experience with Databricks, Delta Lake, Airflow, AWS, and GCP, plus MLOps proficiency in experiment tracking, model monitoring, and reproducible deployment.
• Production experience with vector databases (Pinecone, Weaviate, Milvus, pgvector, or equivalent) and with designing and operating agentic AI systems, including prompt engineering, agent orchestration, tool use, or integrating LLMs into ML pipelines.
• A clear communicator who translates technical work into design docs and cross-functional conversations, and an active mentor who gives direct…