Embedded Data Scientist: Causal Inference & Growth

Socket.dev; New York (NY); On-site; USD 130,000 - 190,000; Full time

Benefits offered by this job

Coaching for creativity

Job summary

Clay is seeking a Data Scientist to join the Data Science and Analytics team. You’ll be durably embedded with a partner team—Product, Marketing, Sales, Finance, or Partnerships—as their data partner, with the deep context to shape what gets built and how the business runs, backed by a central data team that operates as one center of excellence.

This role is for someone who wants to own challenging, ambiguous problems end to end: designing experiments, building predictive models, and designing

Qualifications

• 5+ years in data science with ownership of experimentation or predictive modeling.
• Strong product and business sense; connect analysis to decisions and revenue.
• Strong statistical and experimentation foundations; know when results are trustworthy.
• Expert SQL and Python.
• Experience using AI tools (Claude Code, Cursor) to accelerate analytical work.
• Clear communicator who can influence senior stakeholders.

Responsibilities

• Own causal inference and experimentation; design and analyze experiments.
• Define how we measure the business; shape metric trees and forecasting frameworks.
• Go deep on tough questions; root-cause anomalies and build new analysis frameworks.
• Partner as an embedded expert to leaders and turn questions into rigorous analyses.

Skills

Data science

Experimentation

SQL

Python

Communication

Stakeholder management

Tools

Snowflake

dbt

Hex

Sigma

Dagster

Github

Cursor

Job description

Clay is seeking a Data Scientist to join the Data Science and Analytics team. You’ll be durably embedded with a partner team—Product, Marketing, Sales, Finance, or Partnerships—as their data partner, with the deep context to shape what gets built and how the business runs, backed by a central data team that operates as one center of excellence.

This role is for someone who wants to own challenging, ambiguous problems end to end: designing experiments, building predictive models, and designing