SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Baseten is hiring a Product Data Scientist to establish data-driven decision-making across the company. You'll work directly with Product, Engineering, and GTM teams to define measurement frameworks, strategy, experimentation, and implementation for a technical, usage-based platform serving AI companies like Cursor, Notion, and Writer.
In this foundational, hands-on role, you'll turn ambiguous questions into analyses, forecasts, and experiments that shape product strategy. You'll work with clickstream data, product events, inference telemetry, and observability data to help Baseten make faster decisions about reliability, performance, adoption, and developer experience.
Key responsibilities include: partnering with Product and Engineering to frame critical questions and define success criteria; establishing metrics across activation, adoption, retention, expansion, reliability, and user experience; designing measurement plans and analyzing A/B experiments and controlled rollouts; diagnosing reliability and scaling behavior by joining customer signals with request, replica, deployment, and cluster telemetry; defining the enterprise customer journey and measuring feature adoption; evaluating releases and recovery by tracking traffic shifts, warm-up/drain behavior, and MTTR; and building source-of-truth reporting and self-serve tools with clear recommendations.
You'll need 5+ years in product data science, product analytics, or quantitative roles, ideally supporting developer platforms, APIs, or B2B products. Deep SQL and Python fluency is essential, along with strong statistical judgment and practical experimentation experience including test design and power analysis. You should have hands-on forecasting expertise with ARIMA, Prophet, or comparable time-series methods, plus experience designing medallion data architectures with raw, conformed, and business-ready models. Familiarity with dbt, semantic layers, data ontology, and BI tools like Sigma or Hex is required.
Preferred qualifications include experience with AI/ML infrastructure, model serving, GPU systems, observability for distributed systems, usage-based pricing, APIs, platform unit economics, capacity planning, and model-serving frameworks like vLLM, SGLang, and Dynamo.
