Senior Data Scientist / AI-ML & Anomaly Detection Lead

• Position Type: Part-Time Consultant / Technical Subject Matter Expert
• Target Allocation: 14–18 hours/week average (Note: Workload is highly concentrated around phase deliverables involving rule modernization, anomaly analytics engine build-out, technical validation, and state acceptance testing).
• Location: Remote (U.S. Based) with periodic travel to Tallahassee, FL as required.

• POC Library Modernization: Evaluate, enhance, and modernize the existing 11-rule baseline POC library to support full enterprise scalability.
• Anomaly Engine Design: Architect transaction-centric anomaly logic, defining data features, risk scoring models, tolerance thresholds, and alert prioritization criteria.
• Machine Learning Optimization: Develop, refine, and deploy supervised and unsupervised machine learning algorithms where they add measurable validation value over standard deterministic rules.
• Model Explainability & Traceability: Maintain absolute, non-black-box transparency across all algorithms. Ensure every flagged transaction generates clear, human-readable logic explanations and evidence usable by state auditors or Inspector General investigators.
• Rigorous Validation Testing: Establish comprehensive, audit-ready validation datasets to measure, verify, and document model accuracy, false-positive metrics, and rule reproducibility.
• Independent Rerun Support: Provide complete technical documentation, test scripts, and system logs to enable independent OCIG technical validation teams to successfully rerun all anomaly detection routines.
• Drift & Performance Monitoring: Develop and implement automated pipeline criteria for model scoring transparency, versioning control, feature mapping, and data drift detection.
• Value-Realization Analytics: Engineer standardized formulas and methodologies to compute quantifiable oversight impacts, including potential cost avoidance, financial recoveries, identified risk exposure, and investigative referrals.

• Experience Baseline: 10+ years of comprehensive data analytics and data science experience, with 5+ years of dedicated, hands-on machine learning engineering.
• Government Context: Documented history delivering data science, predictive modeling, or advanced analytics solutions within a federal, state, military, or local government framework.
• Core Tech Stack: Advanced, hands-on mastery of Python and SQL for complex data manipulation and engineering.
• Advanced Analytics Toolkit: Deep expertise across supervised and unsupervised learning, classification, clustering, statistical forecasting, and advanced feature engineering.
• Model Calibration: Proven experience in model evaluation, threshold calibration, and exhaustive false-positive or false-negative impact analysis.
• Investigative Translation: Demonstrated ability to translate raw model outputs into defensible, audit-ready forensic evidence rather than merely outputting an unweighted probability score.
• Scientific Reproducibility: Experience producing comprehensive technical artifacts, configuration baselines, and model documentation sufficient for independent third-party replication.
• Domain Expertise: Strong experience working with highly disparate, transaction-level data sets including financial, procurement, contract management, or purchasing card ledger systems.
• Vetting & Location: Must be a U.S.-based citizen or resident. Must be able to successfully clear an FDLE Level II background screening (including fingerprinting) within 5 business days of contract award.

• Prior experience engineering fraud, waste, abuse, improper payment, financial crime, or corporate risk analytics models.
• Working context with data from oversight agencies such as Treasury, FinCEN, CMS, Census, or state-level Medicaid and revenue departments.
• Hands-on production experience utilizing Azure Databricks, MLflow, or Azure Machine Learning within secure Government Cloud environments.
• Deep familiarity with building hybrid detection architectures that seamlessly blend deterministic business rule repositories with machine learning anomaly detection.
• Prior usage of formal model cards, explainability packages, and structured human-in-the-loop validation review workflows.

1. The Government Customer: Explicitly name the agency and the specific system context (e.g., Treasury, DHS, DLA).
2. The Specific Problem: Detail the exact compliance, oversight, or fraud variant the model was designed to detect.
3. The Technical Approach: Name the specific datasets, algorithms, models, and toolsets utilized (e.g., Clustering, SAS Enterprise Miner, Python).
4. Personal Contribution: Detail exactly what you personally engineered, configured, or deployed regarding features, thresholds, or pipelines.
5. Deployment & Validation Status: State the true production or operational status of the project, including the exact validation method used to prove model accuracy.
6. Quantifiable Outcomes: Provide the measurable results achieved, such as false-positive reduction rates, percentage of pipeline reliability, or volume of cost avoidance unlocked.

“We are an equal opportunity employer and all qualified applicants will receive consideration for employment without regard to race, color, religion, sex, national origin, sexual orientation, gender identity, disability or protected veteran status”

Texting Privacy Policy

• Message type: Informational; you will receive text messages regarding your application and potentially regarding interview scheduling.
• No mobile information will be shared with third parties/affiliates for marketing/promotional purposes.
• Message frequency will vary depending on the application process.Msg & data rates may apply.
• OPT out at any time by texting "Stop".