Relative Geological Time Extraction
Deep learning system for extraction of Relative Geological Time representations from volumetric seismic data.
AI Systems for Volumetric Scientific Data
Applied AI for scientific, spatial, and volumetric data - including seismic, geospatial, and other 3D domains.
Deep learning, scientific computing and large-scale volumetric processing.
AI Engineer / Applied Scientist with a PhD in Computer Science specializing in deep learning for large-scale volumetric and spatial datasets.
Experience spans 3D deep learning, structural segmentation, and high-performance scientific computing - built primarily through seismic interpretation systems, with methods directly transferable to other volumetric and spatial domains such as remote sensing, subsurface energy monitoring, and scientific or medical imaging.
Technical Capabilities
Specialized in machine learning systems for scientific and spatial data processing, combining research-oriented algorithm development with production software engineering.
Selected Technical Work
All work below was developed during previous commercial engagements and is shared here with the employer's permission.
Fault Segmentation
3D deep learning system for automatic fault detection and structural segmentation in volumetric seismic data. Deployment included ONNX-based inference and integration into a CUDA-accelerated C# application.
Synthetic Data & Validation
Development of synthetic geological data generation methods used for validation, controlled experiments, and robustness analysis of deep learning models operating on seismic volumes.
The video presents Relative Geological Time extraction on a synthetic validation dataset generated for controlled experiments and model evaluation.
Angular Unconformity Detection
Experimental deep learning approach for detection of angular unconformities in volumetric seismic data. Focused on identification of large-scale structural discontinuities in real seismic datasets.
Research Background
PhD research focused on local learning algorithms and geometric partitioning methods for high-dimensional feature spaces, including implementation of the proposed learning framework and algorithms in C#.
Collaboration & Contact
Interested in technically challenging contract-based collaborations where machine learning is applied to solve real scientific and engineering problems.
Particularly interested in projects involving space technologies, earth observation, geoscience, subsurface energy and environmental monitoring, medical and scientific imaging, industrial inspection, and other large-scale spatial or volumetric data processing challenges.
Primarily interested in remote contract work with flexible collaboration models.
Email, GitHub, and LinkedIn links are available in the sidebar.