Data Scientist

Pearl is seeking a Data Scientist to lead advanced research and predictive modeling, focusing on analyzing residential housing and energy performance data. The role involves utilizing causal inference and anomaly detection techniques to improve the accuracy of Pearl’s proprietary SCORE models and developing new performance metrics. Additionally, this position acts as the primary technical liaison for external research partners and contributes to the dissemination of findings through white papers and publications

• →Manage research, conducted in partnership with external consultants and statistical firms, that identifies correlations and causal relationships between home performance data and other housing-related data (e.g., energy cost and mortgage performance), using techniques such as regression analysis, propensity score matching, and (where data permit) instrumental variable methods, and ensuring causal claims are supported by appropriate causal inference methods rather than inferred from controlled regression alone.
• →Analyze Pearl's ~92 million residential SCOREs and energy models to identify homes where the SCORE or model output is unlikely to accurately reflect the home's actual physical configuration or energy consumption, using anomaly detection, outlier analysis, and validation against field-collected data on home characteristics.
• →Analyze modeled energy consumption, home physical characteristics, and utility billing data to identify and implement improvements to the predictive accuracy of Pearl's energy models.
• →Analyze relationships between field-collected home performance characteristics and SCORE outputs to identify opportunities to improve SCORE accuracy, using techniques such as feature importance analysis and comparison against field-validated benchmarks.
• →Support development of new performance metrics (e.g., Total Cost of Ownership) by identifying and validating relevant data sources and analytical approaches.
• →Evaluate opportunities to integrate climate risk data into the SCORE, to improve predictive precision around homes’ climate vulnerability, and to analyse the relationships between homes’ resilience features and ability to withstand extreme climate events.
• →Serve as the primary technical point of contact for external data and statistical partners.
• →Assist with the authorship of white papers, briefs, and other publications documenting the research described above for publication on Pearl’s research page, academic journals, etc.

• • Master's degree in Statistics, Economics, Data Science, Applied Mathematics, or a related quantitative field (or equivalent experience)
  • 4+ years of applied experience in statistical analysis and predictive modeling, ideally involving large, real-world (non-experimental) datasets
  • Demonstrated hands-on experience with causal inference methods — regression analysis, propensity score matching, and instrumental variable approaches — and a clear understanding of when correlation-based methods are and are not sufficient to support causal claims
  • Experience with anomaly detection and outlier analysis techniques applied to large datasets
  • Strong proficiency in a statistical/analytical programming language (Python or R) and SQL
  • Experience validating model outputs against ground-truth or field-collected data
  • Ability to translate statistical findings into clear, non-technical explanations for internal stakeholders and external partners
  • Experience working directly with external consultants, research firms, or academic partners on collaborative analytical projects

• • Experience with feature importance analysis and model interpretability techniques
  • Familiarity with housing, real estate, energy, or utility data (assessor records, permit data, utility billing, energy modeling)
  • Experience integrating or evaluating climate/environmental risk data into predictive models
  • A track record of authoring or co-authoring published research
  • Experience working with ambiguity and scale - large datasets, real-world conditions, innovative methodology
  • Comfortable working semi-independently, with support and partnerships