
ML4PEGs
Course designer and instructor
Machine Learning for Petroleum Engineers and Geoscientists
- Institution
- Open course
- Period
- June 2026 – October 2026
- Teaching assistant:
- Haambayi Mudenda
A course designed and taught to bring machine learning methods to petroleum engineering and geoscience problems, from Python fundamentals to deep learning and physics-informed neural networks, with well-log, seismic and reservoir examples throughout and a capstone project.
Course outline23 lessons
Month 1: Python and data foundations
- Course overview and environment setup
- Python basics: variables, lists and mathematical operations
- NumPy and well-log data
- Python essentials and Matplotlib: plotting, control flow and functions
- While loops, functions, lambda and exception handling
- Pandas and tabular well data: filtering, grouping and LAS files
- Advanced Pandas, plotting and multi-track well logs
- Introduction to machine learning: the ML workflow and a first model
Month 2: Classical machine learning
- Classification: predicting formation type from well logs with decision trees
- Model improvement: cross-validation and random forests
- Unsupervised learning: clustering and PCA for reservoir zonation
- Model optimization: hyperparameter tuning and explainable AI
Month 3: Deep learning
- Neural network intuition: from biological neurons to a first network
- Deep learning with PyTorch: tensors, autograd and a training loop from scratch
- Convolutional neural networks: seismic fault detection
- Deeper CNNs: regularization, evaluation metrics and feature maps
- Recurrent neural networks: forecasting reservoir pressure
- LSTMs: gated memory for time series
Month 4: Scientific ML and capstone
- Physics-informed neural networks: the diffusivity equation from scratch
- Capstone, part 1: scoping, data and an honest baseline
- Capstone, part 2: modeling, error analysis, uncertainty and delivery
Bridge lessons
- Git and GitHub fundamentals: version control and an ML portfolio
- Git and GitHub collaboration: branches, merge conflicts and teamwork
Student reviews
No reviews yet. If you took ML4PEGs, I would value your feedback.