Senior Data Scientist, Product Analytics

How You’ll Contribute:

• Advanced Analytics: Conduct comprehensive analysis of product usage, user behavior, and performance metrics to identify opportunities for improvement

• Experimentation Design & Management: Design, implement, and manage A/B tests, causal analysis, and other experiments. Define clear hypotheses, target segments, and success metrics

• Experimentation Analysis: Analyze experiment results using statistical methods, providing detailed reports and recommendations. Identify statistically significant findings and translate them into actionable insights

• Integration Measurement: Own the measurement and experimentation strategy for the embedded DFM AI + IQE integration with our partner. Define success metrics and instrumentation across the in-CAD experience (Solid Edge, NX, Designcenter) and the Xometry marketplace, design cross-surface experiments and causal analyses that join the two funnels, work through the identity-resolution, attribution, and statistical-power challenges of low-volume enterprise cohorts, and quantify the funnel from designer intent in Teamcenter to manufactured part — making the digital thread’s business impact legible to leadership

• KPI Definition & Tracking: Define key performance indicators (KPIs) and establish robust tracking and reporting mechanisms, including KPIs that make sense in a PLM/CAD context (e.g., time-from-design-to-quote, in-CAD-to-order conversion, designer adoption per seat) as well as in the marketplace

• Dashboarding & Visualization: Develop and maintain dashboards to visualize experiment results, key metrics, and trends, making data accessible and understandable to stakeholders

• Collaboration: Work closely with product managers, engineers, designers, machine learning scientists, and — for the Siemens integration — partner engineering teams to implement and iterate on experiment findings and product improvements

• Data Storytelling: Communicate complex data insights and experiment results clearly and compellingly to diverse audiences, including leadership

• Data Integrity: Ensure data accuracy and integrity for all analysis and experimentation, including the cross-system data flows that span the Siemens and Xometry surfaces

• Methodology Development: Contribute to the development and improvement of our experimentation methodologies and best practices, especially for partner-integrated and multi-surface products

What You'll Bring to Xometry:

• Bachelor’s degree in Computer Science, Computer Engineering, Data Analytics, Mathematics, Statistics, Information Systems, Economics, or another quantitative discipline

• 6–8 years of relevant experience

• Strong proficiency in data analysis, statistical modeling, and data visualization

• Proven experience in designing, implementing, and analyzing A/B tests and other experiments

• Solid understanding of statistical significance and hypothesis testing

• SQL is a must. Python is nice to have. Any experience with Looker is also desirable

• Knowledge of data science concepts and methodologies

• Excellent communication and presentation skills, with the ability to explain complex data insights — including to external partner engineering teams

• Strong analytical and problem-solving skills

• Experience working in a product-driven environment

• Attention to detail and a commitment to data accuracy

Nice to have:

• multi-surface / partner-integration analytics: Experience designing experiments and metrics across multi-surface or partner-integrated products (embedded experiences, CAD/PLM integrations, B2B SaaS plug-ins, extension or in-app funnels), including the identity-resolution, attribution, and statistical-power problems that come with low-volume enterprise cohorts

• CAD/PLM/manufacturing fluency: Familiarity with CAD, PLM, or manufacturing workflows; comfort talking with designers, manufacturing engineers, and partner engineering teams about how they actually use tools day-to-day

• causal methods at scale: Hands-on experience with quasi-experimental and causal inference methods (difference-in-differences, synthetic control, CUPED, switchback, geo-experiments) for situations where clean A/B isn’t feasible — common when the experimental unit is an enterprise account or a designer inside a partner environment