Note: The job is a remote job and is open to candidates in USA. General Dynamics Information Technology (GDIT) is a global technology and professional services company delivering technology solutions and mission services to government, defense and intelligence organizations. The Senior Data Scientist will build, validate and deploy supervised machine learning models for healthcare fraud, waste and abuse analytics, while establishing machine learning and model production practices for the program.
Responsibilities
- Designing, training and validating supervised models that score providers and billing patterns for FWA risk, using investigative case-level data, payer feedback on referred leads, and public exclusion and enforcement data as labels, including the entity resolution to link enforcement records to providers in claims
- Designing validation for the actual conditions: labels lagging billing behavior by years, coverage limited to leads previously referred, extreme class imbalance, and schemes that shift faster than confirmation arrives
- Engineering features against billions of claim records within the warehouse rather than extracting data to local memory, using Python and SQL, alongside data engineers and Business Intelligence Developers
- Delivering output that supports action. Investigators need the specific claims, the pattern and the basis for the finding, so each model carries a human-readable rationale and claim-level evidence alongside the score, adjusted for case mix and specialty and ranked so that precision at the top of the review queue is the operative measure
- Deploying models into production and keeping them healthy, including scheduled execution, versioning and drift monitoring, and establishing the modeling and deployment practices the Data Science team adopts going forward
- Collaborating with FWA Subject Matter Experts to separate genuine anomalies from patterns explained by coverage policy or claim edits, and communicating methodology and limitations to HFPP Partners and stakeholders so that output is adopted and acted upon
Skills
- Amazon Web Services (AWS), Healthcare Claims, Predictive Modeling, Python (Programming Language), Supervised Learning
- 5 + years of related experience
- * Master's in a quantitative field (statistics, computer science, engineering, applied mathematics or related), or a Bachelor's with equivalent hands-on experience
- * 5+ years building, validating and delivering supervised machine learning models on real-world data, including work in which labels were incomplete, delayed or biased
- * Experience deploying models into production and maintaining them: scheduling execution, versioning and drift monitoring, with data engineers
- * Python and SQL, including feature engineering within the data warehouse at very large scale rather than extracting to a local environment
- * 2+ years working with healthcare claims data (Medicare, Medicaid or commercial) and coding systems (e.g., ICD-10, CPT, HCPCS, DRG)
- * Experience with validation design for imbalanced, temporally shifting problems: out-of-time evaluation, leakage detection, calibration and precision-focused metrics over ranked output
- * Ability to explain model output to a non-technical investigator, defend methodology to technical audiences, and present analytic outcomes to clients and stakeholders
- * Must be able to obtain/maintain Public Trust
- Work visa sponsorship will not be provided for this position
- Candidates must reside in the United States
- * Graph or network analytics, entity resolution and record linkage for identifying collusive relationships across payers
- * Positive-unlabeled, semi-supervised or active learning against a capacity-constrained review queue
- * Modeling in a regulated or adverse-action setting where explainability and fairness were requirements
- * Anomaly detection, peer-group construction and case-mix methods (e.g., HCC); AWS and/or Snowflake, including Snowpark or model lifecycle tooling
- * Healthcare FWA or program integrity datamining in multi-payer databases; payer coverage policy (LCDs, NCDs) and claim edits (e.g., NCCI)
Benefits
- Remote role; candidates must reside in the United States.
- Medical plan options, some with Health Savings Accounts
- Dental plan options
- Vision plan
- 401(k) plan with the ability to contribute both pre- and post-tax dollars up to IRS annual limits and receive a company match
- Full flex work weeks where possible
- Paid time off plans, including vacation, sick and personal time, holidays, paid parental, military, bereavement and jury duty leave
- Typically 15 days of paid leave per calendar year for vacations, personal business, and illness
- An additional 10 paid holidays per year
- Paid Family Leave of up to 160 hours in a rolling 12 month period for eligible employees
- Short- and long-term disability benefits
- Life insurance
- Accidental death and dismemberment insurance
- Personal accident insurance
- Critical illness insurance
- Business travel and accident insurance
Company Overview