Postdoc: Deep Learning for Subsurface Permeability

University of Pittsburgh; Pittsburgh (Allegheny County); On-site; USD 60,000 - 73,000; Full time

Job summary

The University of Pittsburgh invites applications for a Post-doctoral Researcher in Machine Learning for Subsurface Multiscale Structure and Characterization, including Permeability, funded under the SMART initiative. This three-year appointment focuses on developing AI/ML tools to estimate permeability from geophysical logs using national lab data.

You will collaborate with NETL scientists and publish in high-impact journals, training deep learning models (CNNs, PINNs, GANs) and deploying them

Qualifications

• PhD completed within the last five years from the start date of the appointment.
• Must be a United States Citizen.
• Experience applying ML/DL to scientific problems.
• Proficiency in Python and ML libraries (TensorFlow, PyTorch).
• Strong written and oral communication skills and ability to work both independently and in a team.

Responsibilities

• Develop ML models to predict rock permeability from ultrasonic data.
• Build ML-ready petrophysical databases from NETL data archives.
• Adapt and retrain PhaseNet to automate P and S wave picking in laboratory data.
• Develop GANs to generate synthetic core data for robust model training.
• Integrate models with wireline logs and potential new core samples for validation.
• Publish findings in high-impact journals and present at major conferences.

Skills

Python programming

TensorFlow

PyTorch

Research & analytical skills

Education

PhD in Geophysics, Geology, Petroleum Engineering, Computer Science, or related field

Tools

PhaseNet

CNNs

PINNs

GANs

Job description

The University of Pittsburgh invites applications for a Post-doctoral Researcher in Machine Learning for Subsurface Multiscale Structure and Characterization, including Permeability, funded under the SMART initiative. This three-year appointment focuses on developing AI/ML tools to estimate permeability from geophysical logs using national lab data.

You will collaborate with NETL scientists and publish in high-impact journals, training deep learning models (CNNs, PINNs, GANs) and deploying them