Remote Lead Data Scientist: Interpretable Scoring

Bigbear.ai; McLean (VA); Hybrid; USD 150,000 - 210,000; Full time

Job summary

BigBear.ai in the United States is seeking a Data Scientist to design and build the v1 rule-weighted scoring logic that converts normalized risk signals into a transparent, defensible score, with explainability preserved for adjudicator workflows and audit needs.

This fully remote role requires Top Secret clearance and travel within the DMV area. You will collaborate with data engineering and application teams to productionize scoring, validate models, and document methodology.

Qualifications

• Bachelor's or higher in a technical field and several years of applied data science experience.
• Experience delivering scoring, ranking, or decision-support models.
• Experience implementing interpretable approaches (rule-based or SHAP explanations).
• Strong Python skills with scikit-learn and data science workflows.

Responsibilities

• Design v1 rule-weighted scoring logic from normalized risk signals.
• Define scoring framework components: features, weights, thresholds, guardrails.
• Create interpretable explanations for scores and drivers for adjudicator reviews.
• Design architecture to evolve from rules to interpretable ML models with auditability.
• Collaborate with data engineering and application teams to productionize scoring.

Skills

Strong Python skills

Explainability/interpretability

Stakeholder communication

Education

Bachelor's Degree

Master's Degree

PhD

Tools

Python

scikit-learn

SHAP

SQL

AWS SageMaker

Lambda

Glue

Neptune

Jupyter

Job description

BigBear.ai in the United States is seeking a Data Scientist to design and build the v1 rule-weighted scoring logic that converts normalized risk signals into a transparent, defensible score, with explainability preserved for adjudicator workflows and audit needs.

This fully remote role requires Top Secret clearance and travel within the DMV area. You will collaborate with data engineering and application teams to productionize scoring, validate models, and document methodology.