Senior Data Scientist — Data Quality & Statistical Methodology

BLN24•Mclean, VA

13d•Hybrid

About The Position

BLN24 is seeking a Senior Data Scientist with a strong statistical-methodology focus to support a large-scale enterprise data and analytics platform modernization effort. This role sits at the intersection of statistical methodology, data-quality measurement, and large-scale data engineering — designing the metrics and methods that determine whether an organization’s core data products can be trusted. A central challenge is completeness: source data is rarely perfect, and values are routinely missing, partial, or unreliable. The ideal candidate understands how external and secondary data sources can be used to responsibly fill those gaps, and can design defensible, statistically sound quality metrics that measure how well the resulting data reflects reality. The platform’s anticipated foundation involves a modern lakehouse/cloud data architecture handling very large datasets from multiple providers. The successful candidate will help define the quality-metric framework and gap-filling methodology for a generation of stakeholders moving off legacy tools and fragmented, manually validated processes.

Requirements

5+ years of applied experience in statistical methodology, data quality, or quantitative research roles
Demonstrated experience with missing-data and imputation methods (e.g., model-based imputation, hot-deck, sequential regression) for filling incomplete data
Experience designing and validating quantitative metrics for decision-support, including an understanding of bias, variance, and false-positive/false-negative trade-offs
Working knowledge of record linkage / entity resolution concepts, sufficient to build sound metrics on top of matched data
Experience with very large datasets on distributed-compute platforms (e.g., Spark-based / lakehouse environments) and strong SQL
Strong proficiency in Python and R
Comfort working with regulated or restricted data and the governance constraints that accompany it
Strong communication skills and the ability to explain and defend methodology to leadership and non-technical stakeholders

Nice To Haves

Master’s or PhD in Statistics, Applied Mathematics, Econometrics, Data Science, or a related quantitative field (a purely software-focused background is not sufficient for the methodology components of this role)
Prior experience supporting large-scale enterprise data programs or platform modernization efforts
Experience using external or secondary data to supplement or complete primary datasets
Familiarity with Databricks and modern lakehouse architectures
Exposure to privacy-preserving analytics techniques or working with regulated data
Ability to read and reconcile legacy statistical codebases (e.g., SAS) alongside modern Python/R workflows
Background in requirements gathering for enterprise data platforms
Experience benchmarking estimates against authoritative reference datasets for anomaly or drift detection

Responsibilities

Design, define, and validate data-quality metrics for very large datasets — not simply reporting numbers, but establishing what each metric means, how it is calculated, and why it is statistically defensible to leadership
Develop and document methodology for filling gaps where source data is missing, partial, or unreliable, using external and secondary reference data, including model-based and imputation approaches
Establish benchmarking approaches that compare data products against authoritative historical and modeled reference datasets to detect drift, bias, and anomalies
Specify the data the platform must ingest to support quality monitoring, and define the checks that flag when an upstream-produced data product looks wrong
Partner with subject matter experts (SMEs) and stakeholders to translate operational and analytical questions into concrete, measurable quality requirements
Work with data engineers to ensure metrics and gap-filling logic run reliably at scale on very large, multi-source datasets built on common keys and governed definitions
Account for data sensitivity throughout, ensuring appropriate aggregation, access controls, and privacy-preserving techniques are reflected in any metric or derived data product
Document methodology and requirements in structured, reusable formats (e.g., requirements matrices and detailed requirement specifications)
Iterate across multiple review cycles with SMEs and fellow methodologists, given the program’s phased, multi-year rollout