Staff Data Scientist, Forecasting

Pinterest Pinterest · Consumer · San Francisco, CA · Core Engineering

Staff Data Scientist, Forecasting at Pinterest. This role involves leading the strategy and implementation of forecasting models for key company metrics, owning the full modeling lifecycle from problem framing to deployment and monitoring. The role requires technical leadership, mentoring junior data scientists, and translating forecasts into business decisions. It focuses on shipping production time-series models with web-scale data.

What you'd actually do

  1. Be the technical lead for the forecasting team. Own the strategy and implementation of forecasting models of key company metrics (e.g., monthly active users), delivering accurate, interpretable forecasts at scale.
  2. Lead the full modeling lifecycle end to end: problem framing, feature engineering, model development and prototyping, experimentation and backtesting, deployment, monitoring/drift detection, and explainability.
  3. Set the forecasting technical vision. Define model architectures and standards, and partner with Engineering to shape the forecasting platform for efficient training/inference today and the scalability needed for the next generation of models.
  4. Translate forecasts into decisions. Present outputs, scenario analyses, and recommendation frameworks to senior leadership with clarity and brevity. This is a high‑visibility role with regular VP-level exposure.
  5. Drive broader time‑series impact beyond point forecasts—e.g., anomaly detection, automated root‑cause analysis, campaign/channel attribution, and early‑warning signals for business health.

Skills

Required

  • 8+ years of combined post-graduate academic and industry experience building and shipping production time‑series/forecasting models with web‑scale data.
  • Bachelor’s degree in a relevant field such as Computer Science or equivalent experience.
  • A track record of delivering adjustable, well‑calibrated, and explainable forecasting systems that informing decision-making.
  • Strong background in time‑series modeling and applied statistics/econometrics
  • Expertise in at least one scripting language (ideally Python).
  • Strong SQL skills (Hive/Presto/Spark SQL) and experience building reliable data pipelines/workflows (e.g., Airflow).
  • Business acumen and ownership mindset
  • Excellent communication skills
  • Proven technical leadership

Nice to have

  • advanced degree (MS or PhD) preferred

What the JD emphasized

  • production time‑series/forecasting models
  • adjustable, well‑calibrated, and explainable forecasting systems
  • technical leadership

Other signals

  • forecasting models
  • time-series
  • production models
  • web-scale data
  • technical leadership