SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Salary: USD 265,000 - 320,000 / annual
Anthropic is seeking a Data Scientist to join the Finance Analytics and Business Intelligence team as a founding member of the Finance Forecasting group. This is a high-impact role focused on building production-grade revenue forecasting models and causal measurement capabilities that drive capacity planning, board reporting, and strategic decision-making.
You will own a core piece of the forecasting program, choosing to lead either production revenue forecasts or causal measurement work. On the forecasting side, you'll scope, develop, backtest, deploy, and monitor models that are visible across the company and directly influence executive decisions. On the causal side, you'll build measurement programs that quantify the true impact of launches and events, replacing guesswork with rigorous, repeatable analysis.
Key responsibilities include: owning end-to-end modeling work (scoping through deployment); building and maintaining backtesting and accuracy-tracking discipline with public model scoring; contributing to research on event-aware architectures, hierarchical reconciliation, and causal designs; translating model outputs into clear recommendations for Finance and leadership; and partnering with Analytics Engineers on feature pipelines and model deployment infrastructure.
You must have substantial production experience with data science, forecasting, or quantitative finance—not just notebook work. Deep fluency in Python and SQL is required, along with strong applied statistics foundations. Depth in either production time-series methods (Prophet, ETS, ARIMA, gradient boosting, neural forecasting, hierarchical reconciliation) or causal inference (difference-in-differences, synthetic control, Bayesian structural time series, event studies) is essential. You should have built backtesting discipline before, be comfortable with public model scoring, and have presented forecasts or causal estimates to executives. A bias for action and comfort in ambiguous, early-stage environments is critical.
Preferred experience includes exogenous-regressor or event-aware forecasting, hybrid or foundation-model forecasting systems (TimeGPT-class), pricing/elasticity or marketing-mix modeling, and forecasting consumption-based or usage-billed businesses (cloud, API, marketplace).
