SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Outtake is building an agentic AI platform that detects, monitors, and removes impersonators and fraud at scale. Founded by ex-Palantir, ex-CTO/founders, and ex-Notion engineers, the company is backed by top-tier investors and operates with strong financial footing.
As Founding Applied Data Scientist, you will define how Outtake understands, measures, and improves product performance, business outcomes, and AI system behavior. You'll build the data foundation that transforms product usage, customer outcomes, model behavior, and business operations into clear decisions and durable systems.
Key responsibilities include:
- Own the analytical data pipeline and infrastructure across product, business, and AI performance data
- Define and maintain the semantic layer in the product analytics stack, including Hex and warehouse models
- Build canonical performance metrics for product and business (activation, usage, retention, customer value, operational efficiency, agent effectiveness)
- Partner with Product and Finance on pricing models, usage-based packaging, margin analysis, and customer-level profitability
- Work with Platform Engineering on AI performance metrics, evals, benchmarking, observability, and reliability reporting
- Design dashboards, analyses, and decision-support systems for fast, high-confidence decisions
- Build data quality checks, documentation, and metric definitions
- Translate ambiguous questions into rigorous analyses and practical recommendations
- Hire and onboard Outtake's Data Team; set long-term roadmap for data infrastructure
Required: 5+ years as Engineer, Analytics Engineer, Data Scientist, or related role. Strong SQL proficiency (Postgres preferred), experience building reliable data pipelines and semantic layers, ability to define product/business metrics from first principles, familiarity with modern AI eval frameworks and LLM observability, strong product judgment, cross-functional collaboration skills, and ability to communicate complex analyses clearly.
Nice-to-have: Hex, Langfuse, Braintrust, Arize, Phoenix, OpenTelemetry experience; ClickHouse, BigQuery, Snowflake, Databricks, DuckDB, dbt; usage-based pricing and unit economics; AI agent/fraud/abuse detection metrics; experimentation and causal inference; internal tools/notebooks; early-stage startup experience.
Outtake offers 100% company-paid medical, dental, and vision; flexible PTO; annual retreats; in-person collaboration in a Brooklyn waterfront office (5 days/week with flexibility).
