Principal Data Scientist, Ads

Reddit Reddit · Consumer · United States · Remote · Ads Data Science

Principal Data Scientist for Reddit's Ads team, focusing on advanced quantitative methods, machine learning, and causal inference to optimize the ads marketplace. This role involves defining and implementing next-generation data science solutions, strategic leadership, and establishing scientific standards for modeling and experimentation.

What you'd actually do

  1. Define the future of Ads Data Science: Own the design and long-term evolution of our core Ads Data Science solutions and infrastructure, building for the next several years of continued scale, revenue growth, and relevance.
  2. Strategic leadership & scientific roadmap: Identify fundamental gaps and opportunities in our current systems (including platform, auction, targeting, and full-funnel relevance models). Lead the strategic design and scientific roadmap for new solutions to significantly improve Advertiser ROI, User experience, and Reddit revenue.
  3. Drive end-to-end impact: Take end-to-end ownership of complex problem domains such as full funnel acceleration, advertiser lifetime value (LTV), and developing advanced predictive and causal frameworks that power Reddit’s strategic investments.
  4. Establish scientific standards: Define and codify best practices for large-scale statistical modeling, economic analysis, causal inference, offline model evaluation, and A/B experimentation to ensure scientific rigor and trust across the Ads data science and engineering teams.
  5. Technical deep dive & authority: Be the undisputed domain authority for data scientists, engineers, and product managers on complex problems involving large-scale behavioral data, foundational models for strategic insights, and the economics of marketplace dynamics.

Skills

Required

  • Python or R
  • SQL
  • statistical analysis
  • economic modeling
  • foundational machine learning
  • optimization techniques
  • experimental design
  • causal inference
  • A/B testing methodologies
  • ads marketplace understanding
  • auctioning/bidding
  • ads creative & format evaluation
  • measurement & experimentation at scale

Nice to have

  • MLOps
  • deep learning
  • reinforcement learning

What the JD emphasized

  • advanced quantitative methods
  • machine learning
  • statistics
  • econometrics
  • quantitative methods
  • measurement
  • prediction
  • optimization
  • statistical modeling
  • economic analysis
  • causal inference
  • offline model evaluation
  • A/B experimentation
  • large-scale behavioral data
  • foundational models
  • marketplace dynamics
  • advanced predictive and causal frameworks

Other signals

  • optimize Reddit's Ads marketplace
  • foundational data science solutions
  • advanced quantitative methods
  • measurement, prediction, and optimization
  • statistical modeling, economic analysis, causal inference