Lead Data Scientist PyTorch

We're looking for a Lead Data Scientist skilled in PyTorch to collaborate with a multidisciplinary team of Geoscientists, Geophysicists, Data Scientists, and Machine Learning Engineers. The role involves translating technical and business requirements into reliable, high-impact deep learning solutions for seismic data, while also enhancing existing seismic ML tools and uncovering new business applications.

• Build and apply deep learning models (U-Nets, CNNs, and related architectures) to denoise and structurally interpret 2D and 3D seismic data
• Enhance the accuracy, robustness, and performance of current seismic ML tools through retraining, architecture improvements, and hyperparameter optimization
• Create models for structural interpretation needs, including fault detection, horizon tracking, and seismic facies classification
• Collaborate with Geoscientists and Geophysicists to validate outputs against expert interpretations and well data, incorporating their input into model refinements
• Spot, assess, and recommend new business opportunities where computer vision and deep learning could benefit exploration and production workflows
• Keep pace with advances in computer vision and seismic ML research, evaluating new techniques for real-world application
• Maintain sound engineering standards for reproducible experiments, model versioning, and production deployment

• A minimum of 5 years in comparable roles (ML Engineer, Data Scientist, or Geophysicist with strong ML expertise), with practical experience applying deep learning to real-world image or volumetric datasets
• Solid grasp of Computer Vision concepts, such as image segmentation, denoising, feature extraction, and convolutional architectures
• Strong command of PyTorch for constructing, training, and assessing deep learning models
• Deep knowledge of U-Net, CNN, and similar architectures applied to 2D and 3D seismic datasets
• Experience with in the Upstream Oil & Gas sector, collaborating with Geoscience teams and seismic datasets
• Familiarity with seismic data fundamentals, including acquisition and processing principles, seismic attributes, noise characteristics, and structural elements like faults and horizons
• Excellent Python programming abilities
• Nice to have Exposure to transformer-based or self-supervised learning models for vision or seismic applications
• Knowledge of MLOps platforms (such as MLflow), Docker, and cloud-based AI services (Azure, AWS, or GCP)