Overview
Own end-to-end product and growth analytics for AI-driven software creation by turning event, unstructured text, and agent traces into predictive, queryable signals—then embedding those insights into product and growth workflows.
What you'll do
- Transform agent trajectories, support tickets, logs, and user prompts into structured, queryable signal using summarize-then-embed-then-cluster pipelines with LLMs.
- Surface early indicators of confusion, bug waves, churn risk, and fraud, and route them to the right team.
- Build predictive models to forecast conversion, retention, expansion, and churn and embed those signals into product and growth workflows.
- Own marketing attribution and MMM by building media-mix and incrementality models for signups and paid conversions.
- Run analytics and A/B/growth tests across the self-serve funnel, web, and mobile; analyze with power/novelty/interference and causal methods.
- Cluster hundreds of thousands of agent trajectories to discover recurring build attempts and failure modes, producing a taxonomy tied to churn prediction.
- Build models for onboarding “aha moment,” gross-margin attribution down to app/cohort, and fraud ring untangling.
What you'll need
- 2 to 5 years in data science or applied ML with focus on product analytics, growth, or user behavior.
- Strong SQL with comfort working with large-scale, event-level behavioral data.
- Classical ML foundations including clustering, embeddings/vector similarity, dimensionality reduction, and classification.
- Experience turning unstructured natural-language data (LLM traces, logs, tickets, free text) into predictive, queryable signal.
- Proficiency in Python and standard data-science tools (pandas, scikit-learn, statsmodels, numpy).
- Data engineering competence to design and ship ETL and data models (dbt or equivalent).
- Experience designing and analyzing experiments with sample sizing, power, significance, novelty effects, interference, and causal methods.
Nice to have
- Marketing Mix Modeling (MMM), media attribution, or incrementality/geo-testing experience.
- Experience at a PLG company with a self-serve funnel and freemium or usage-based pricing.
- Modern data stack/tools such as BigQuery and dbt, and product analytics platforms like PostHog, Amplitude, Mixpanel, or Segment.
- Causal inference methods such as difference-in-differences, synthetic control, and propensity score matching.
- Experience in fraud, trust and safety, or abuse analytics.
