- Design and build standalone quantitative models for forecasting, optimization, scoring, pattern detection, and relationship analysis
- Develop model components with clear inputs, outputs, assumptions, performance expectations, and testing criteria
- Build and maintain data pipelines that combine historical data, vendor inputs, market signals, trend data, and other enterprise data sources
- Prepare model-ready features from incomplete, inconsistent, or unreliable real-world data
- Develop backtesting and evaluation frameworks to compare model outputs against historical outcomes before deployment
- Validate model performance, accuracy, reliability, confidence ranges, and known limitations
- Implement model monitoring for drift, input quality, accuracy changes, and data-source reliability over time
- Register models in a governed model registry with clear versioning, explainability, and audit support
- Define and document callable model interfaces, including inputs, outputs, latency, confidence bounds, and performance characteristics
- Document model design, assumptions, decision logic, limitations, and validation results for client, stakeholder, and audit review
PythonMachine LearningData science+2 more