I live in Trondheim and work as senior data scientist at Aker BP ASA. Previously, I worked as software engineer at Whitson and data scientist at Prores AS, Whitson AS, and Asia Pulp and Paper. I've spent 5+ years working in this role for different industries in oil and gas, pulp and paper, and forestry. My core expertise is computer vision for subsurface core/cutting images, seabed mapping, and aerial drone/LiDAR photogrametry, with interests in LLM, AI engineering, and advanced machine learning models.

I use my GitHub mostly for educational purposes and OSS (open-source software) development. I like to help many university students and professionals to step their feet into Python programming.

I am particularly interested in the use of agentic AI for science and engineering (one of them in geos). Here are my most recent works:

• Boglodite: Agent for seismic interpretation
• Log-to-seismic prediction: Log-to-seismic prediction
• Private agentic RAG system: Private agentic RAG system

• DEUP LLM (Research): Epistemic uncertainty quantification for LLMs using Direct Epistemic Uncertainty Prediction (DEUP) methodology
• SciSandboxAI: Agent for scientific research and experiments
• Knowledge graph from text and image embeddings (Research): Knowledge graph from embeddings for document retrieval

These are several Python libraries that I've developed over 5 years and I use to teach my students:

Library for geophysical data processing, such as gravity data anomaly separation using Fast Fourier Transform / Second Vertical Derivative (FFT/SVD), seismic data attribute calculation, interactive viewer, Amplitude-versus-offset (AVO) modelling, seismic inversion, and rock physics.

Library for reservoir engineering analysis, such as Original-oil-in-place (OOIP) contouring and calculation, Pressure-Volume-Temperature (PVT) correlation, material balance analysis for gas (dry gas and water-drive) and oil reservoirs (saturated and undersaturated), well test simulation and interpretation, and Decline Curve Analysis (DCA).

Library for visualization and analysis of well log data, including triple combo visualization, neutron density plot, and log interpolation.

This repository stores codes, data, and materials of trainings that I give to students. This is a perfect place to learn Python.