Data Scientist

Perplexity is AI for people who expect more. This role brings that standard to how we understand our users, shape our product, and decide what to build next.

• Develop product insights: analyze user behavior to inform product roadmap, accelerate adoption, and identify opportunities to improve the user experience.
• Design and analyze experiments: form hypotheses, define success metrics, run A/B tests, interpret results, and turn findings into product recommendations.
• Define the metrics that matter: build the metrics, guardrails, dashboards, and reporting workflows that help teams understand product and company health.
• Partner across functions: work closely with engineering, product, growth, design, and user research to answer ambiguous questions and drive decisions.
• Build reusable data assets: create tables, models, and documentation that make analysis faster, more consistent, and easier for humans and AI systems to use.
• Use AI to scale data science: automate recurring analysis, build AI-assisted workflows, improve documentation, and turn one-off investigations into repeatable systems.
• Tell the story clearly: communicate findings, assumptions, uncertainty, and recommendations in a way that helps teams make decisions.

• 6+ years of experience as a data scientist or closely related role.
• Strong product sense: you understand user behavior, product tradeoffs, and how to connect analysis to decisions.
• SQL expertise: you can navigate a complex data warehouse on your own, reason about grain and joins, and debug data issues when something looks wrong.
• Experimentation depth: you have significant experience designing, running, and analyzing A/B tests.
• AI-native working style: you use LLMs and AI tools to move faster without outsourcing analytical judgment.
• Metrics and dashboard experience: you've built useful reporting in tools such as Omni, Mode, Hex, Looker, or similar, and know how to turn metrics into better product decisions.
• Comfort with ambiguity: you can take an open-ended question, structure it, execute the analysis, and make a clear recommendation.
• End-to-end ownership: you take responsibility for the work from problem definition to stakeholder adoption.

• Define data science at Perplexity: you'll help set the bar for how an AI-native data science team operates.
• Work on questions with no playbook: AI-native products are too new for established benchmarks. What engagement, retention, and user success look like for AI products that answer questions and get real work done is still an open question, and you'll help define it.
• Use AI where it actually matters: you'll use frontier tools to make analysis faster, more repeatable, and more useful.
• Direct impact: small team, high ownership, and work that shapes product and company direction.