Machine Learning Engineer
- Budapest
- Állandó
- Teljes munkaidő
- Architect, build, and productionize end-to-end machine learning systems for large-scale demand forecasting across global prestige beauty brands, covering consumer POS and customer order signals.
- Develop modular, reusable, and scalable ML pipelines for time series forecasting (e.g., hierarchical models, deep learning models, probabilistic ensembles) leveraging state-of-the-art libraries and in-house frameworks.
- Own model lifecycle management: automated training, hyperparameter tuning (e.g., Optuna/Bayes), backtesting, CI/CD, model versioning (e.g., MLflow), and monitoring using best-in-class MLOps practices.
- Engineer robust data pipelines across distributed compute platforms (e.g., PySpark on Databricks, Delta Lake, Ray) to support feature generation, near real-time scoring, and model retraining at scale.
- Integrate ML forecast systems into downstream supply chain platforms (e.g., O9, SAP IBP), ensuring API-based delivery, schema evolution handling, and traceability.
- Design automated model evaluation frameworks with multi-metric validation (e.g., sMAPE, MASE, forecast bias/stability) across business units, regions, and time horizons.
- Collaborate closely with software engineers, data scientists, product managers, and global planners to productionize research and align modeling strategies with business objectives.
- Programming & Tooling: Expert in Python (NumPy, pandas, scikit-learn, PyTorch/TF), PySpark, MLflow, Git, Docker, Airflow.
- Forecasting & Modeling: Deep expertise in classical/statistical forecasting (ETS, ARIMA, Exponential Smoothing), machine learning models (XGBoost, CatBoost), and deep learning models (LSTM, N-BEATS, Temporal Fusion Transformer).
- MLOps & Infrastructure: Experience with CI/CD for ML (e.g., using Azure DevOps or GitHub Actions), container orchestration (e.g., Docker, Kubernetes), model monitoring (e.g., Prometheus/Grafana), and experiment tracking.
- Scalability & Performance: Skilled in building distributed ML pipelines using Spark and Delta Lake, with focus on low-latency inference and scalable retraining workflows.
- Data Engineering: Proficient in ETL/ELT workflows, schema design, and pipeline optimization; experienced with event-driven data and batch/stream processing.
- System Design & Architecture: Capable of designing resilient ML systems that integrate with enterprise-scale supply planning tools and forecasting APIs.
- Communication & Collaboration: Adept at translating complex technical topics into business insights for cross-functional teams, including supply chain, planning, and IT stakeholders.