Data Analysis . Feature Engineering

Project start: As soon as possible / by mutual agreement
Duration: 6+ months (extension possible)
Workload: 100%
Location: Primarily remote, with 9 onsite days per year in Frankfurt am Main for planning sessions

Responsibilities
 

Data Analysis & Feature Engineering
- Perform complex ad-hoc data analyses to identify relevant factors influencing forecasting models
- Prepare, explore, and transform large datasets using Python (e.g., pandas, polars) and SQL
- Develop robust features to improve model accuracy and quality
 

Forecasting & Machine Learning Model Development
- Design, implement, and train statistical, machine learning, and time-series models
- Work with methods such as regression models and neural networks
- Select appropriate modeling approaches based on business requirements, data structure, interpretability, and scalability
 

Model Validation & Quality Measurement
- Define, calculate, and interpret key model evaluation metrics
- Conduct cross-validation, backtesting, and stability analyses to ensure reliable and generalizable forecasts
 

Model Comparison & Performance Monitoring
- Compare different modeling approaches using standardized evaluation frameworks
- Monitor model performance over time, including data drift and model drift detection
- Analyze deviations between forecasts and actual outcomes
 

Operationalization & Integration
- Support integration of forecasting models into production data pipelines on AWS
- Collaborate with Data Engineers and BI Developers to deliver model outputs for reporting, planning, and decision-making
 

Data Quality & Model Robustness
- Ensure high data quality through plausibility checks, outlier detection, and missing-data handling strategies
- Assess model sensitivity to data changes and assumptions
 

Documentation & Communication
- Maintain clear documentation of model assumptions, training processes, and validation results
- Communicate forecast quality, uncertainties, and limitations to business stakeholders and management
 

Continuous Improvement
- Continuously enhance existing models based on new data, stakeholder feedback, and changing business requirements
- Stay updated on best practices and new methods in forecasting and model evaluation
 

Governance, Security & Compliance
- Implement IT security policies, regulatory requirements, and internal programming standards
- Ensure data privacy, security, and compliance in all data processing activities

Required Qualifications
Mandatory Requirements
- 5+ years of practical experience in designing, implementing, and validating forecasting, machine learning, and time-series models in production environments
- 5+ years of hands-on Python experience, especially with pandas or polars for data analysis, feature engineering, and model development
- 2+ years of SQL experience for data querying and preparation
- 3+ years of experience in model validation and quality measurement, including evaluation metrics, cross-validation, and stability analysis

Nice-to-Have Qualifications
- Experience with forecasting, machine learning, or time-series models based on graph databases (minimum 2 projects)
- Experience integrating forecasting models into production AWS data pipelines (minimum 1 project)
- Experience analyzing and processing data from railway control and signaling systems (minimum 1 project)