We’re seeking a Deep Learning Data Scientist to join our research-driven team developing real-time recommendation systems for industrial production environments. You'll be responsible for designing, training, and evaluating deep learning models that help optimize processes, detect anomalies, and uncover hidden patterns insensor and control data.
This is a hands-on role for someone passionate about solving challenging real-world problems using modern machine learning techniques. You'll collaborate closely with other researchers, data engineers, and system developers to bring your models to life.
- Design and implement deep learning models tailored to time-series sensor data, control system behavior, and anomaly detection.
- Develop custom architectures (e.g., RNNs,Transformers, Autoencoders) adapted to noisy, multi-dimensional data streams.
- Conduct thorough experimentation and benchmarking to evaluate model performance and robustness.
- Work alongside data engineers to define and prepare high-quality training datasets from production environments.
- Optimize models for runtime performance and interpretability in critical systems.
- Collaborate with research engineers to integrate DL components into production pipelines.
- Stay current with state-of-the-art researchand assess its applicability to our domain.
- M.Sc. or Ph.D. in Computer Science, Machine Learning, Electrical Engineering, or a related field.
- 3+ years of experience developing deep learning models, ideally for time-series or structured sensor data.
- Proficiency in Python and deep learning frameworks such as PyTorch or TensorFlow.
- Strong understanding of model architectures, training dynamics, overfitting prevention, and evaluation strategies.
- Experience working with noisy or incomplete real-world data.
- Ability to independently run experiments, interpret results, and iterate quickly.
- Familiarity with anomaly detection, predictive maintenance, or industrial control systems.
- Knowledge of signal processing, control theory, or process optimization.
- Experience with MLOps tools (e.g., MLflow, DVC, Weights & Biases).
- Exposure to hybrid model architectures (combining physics-based and data-drivenapproaches).
