Senior Data Scientist, AI Product Insights

About the role

The Senior Data Scientist, AI Product Insights integrates cutting edge data science into Mixpanel's products and serves as a methodological resource for cross functional teams. This role focuses on building the analytical reasoning layer that powers proactive trustworthy insights and agentic experiences for customers. You will design and apply advanced statistical methods to understand behavior and drive business outcomes using rigorous causal inference. The position translates complex statistical concepts into clear narratives for Product and Engineering stakeholders without sacrificing technical precision. You will work at the intersection of product analytics and AI, shaping how customers learn from behavioral event data. Collaboration spans product, engineering, data infrastructure, and finance to ensure analytical solutions are both rigorous and actionable. This role owns the development of survival models, causal impact studies, and forecasting systems that directly inform product strategy. You will partner closely with AI platform, analysis, and data infrastructure teams to operationalize insights into the Mixpanel platform.

Key facts

What you'll do

Design and apply causal inference methods such as propensity score matching, regression discontinuity, difference in differences, or instrumental variables to identify behaviors that drive business outcomes.

Develop survival analysis and retention models, including Cox proportional hazards and Kaplan Meier approaches, that power Signals and Simulation outputs.

Translate statistical concepts, such as propensity scores or survival curves, into plain language for Product and Engineering audiences without losing precision.

Use SQL for data access, exploration, and validation across analytical workflows and production systems.

Evaluate and ground LLM generated explanations or recommendations against statistical outputs, including consistency checks and error analysis.

Use time series forecasting methods, including classical approaches like ARIMA and exponential smoothing, and familiarity with modern foundation models such as TimesFM or Chronos.

Apply clustering and similarity methods to behavioral or user event data to uncover segments and patterns that inform product decisions.

Operationalize offline batch analyses in production systems, working hand in hand with software engineers to integrate models into scalable pipelines.

Perform feature engineering from raw event streams to create robust inputs for modeling and simulation workflows.

Conduct clear written and verbal communication of complex statistical methods to non technical stakeholders, aligning technical work with product decisions.

Comfort working in fast moving product settings where analytical rigor and practical delivery are equally important.

Leverage AI coding tools such as Claude Code or Cursor to accelerate modeling iteration and validation cycles.

Prior experience with large scale behavioral event data in product analytics, growth, or observability domains is valued.

Background in structural equation modeling or causal DAGs for multi metric impact modeling relevant to Simulation is preferred.

Maintain and iterate on models in production, monitoring performance and ensuring ongoing validity as product behavior evolves.

Requirements

Advanced degree in a quantitative field or equivalent experience demonstrating expertise in causal inference.

5+ years applying statistical modeling to real world product or business problems.

Hands on experience with causal inference techniques, including propensity score matching, regression discontinuity, difference in differences, and instrumental variables.

Practical background in survival analysis or retention modeling using methods such as Cox proportional hazards or Kaplan Meier.

Strong Python proficiency across the analytical stack, including statsmodels, scikit learn, pandas, and libraries for survival analysis, clustering, and time series modeling.

Experience with time series forecasting methods, including classical approaches like ARIMA and exponential smoothing, and familiarity with modern foundation models such as TimesFM or Chronos.

Experience with clustering and similarity methods applied to behavioral or user event data.

Clear written and verbal communication skills for explaining complex statistical methods to non technical stakeholders.

SQL fluency for data access, exploration, and validation in product environments.

Comfort working in fast moving product settings where analytical rigor and practical delivery are equally important.

Demonstrated ability to learn quickly and apply new tools, including AI coding assistants, to accelerate modeling and insight generation.

A portfolio or track record of analyses that demonstrate impact on product decisions is highly valued.

Nice to have

Experience with large scale behavioral event data in product analytics, growth, or observability domains.

Familiarity with feature engineering from raw event streams.

Background in structural equation modeling or causal DAGs for multi metric impact modeling relevant to Simulation.

Experience operationalizing offline batch analyses in production systems.

Hands on work directly in production codebases alongside software engineers.

Prior experience at analytics, observability, or growth platforms.

Experience evaluating or grounding LLM generated explanations or recommendations against statistical outputs, including consistency checks.

Comfort using AI coding tools such as Claude Code or Cursor to accelerate modeling iteration.

Practical notes

This role requires a hybrid work arrangement based in San Francisco.

The position may involve on call responsibilities for model behavior and validation.

Travel is not routinely required for this role.