Data Scientist
+10 years of Multidisciplinary Geoscience Experience
Adept at statistical and spatial data analysis, GIS, machine learning, and subsurface modeling
Core Skills
- Data Science: Machine learning algorithms, ML-toolkits: TensorFlow, Keras, Scikit-learn, Statsmodels, Deep Learning, Pattern Recognition and clustering
- Advanced Statistics/ Spatial Analysis: Spatial, statistical, and stochastic modeling, EDA, PCA, FA and discriminant analysis, regression techniques, Generalized Linear Models, Bayesian learning
- GIS: Projections, geocoding techniques, Esri’s GIS tools, spatial analysis, USGS quadrangle maps, aerial/ satellite imagery, Google Earth and scripting using python (arcpy, pyshp, GeoPandas, pyproj, Shapely), QGIS
- Programming and scripting skills: Proficient in Python, MATLAB/Octave and experienced in C++, VBA, familiar with R, Fortran
- Database: EDW (Snowflake) using SQL query and python scripts, Relational Database Management (MS Access, SQLite)
- Data Visualization: Matplotlib, Seaborn, ParaView, Mayavi, VTK, mapping in ArcGIS, Familiar with d3, poltly, Data Studio
- Geostatistical/Geological Modeling: Geomodeling package (Petrel, Rockworks, DataMine), geostatistical Software (ISATIS, SGeMS, GSLIB), Matlab and Python libraries written for FKA, KED, MIK, and Bunch-pasting MPS.
- Natural Resources: formation, data acquisition, interpretation, characterization, modeling (StreamSim, Modflow)
- Pursuing: CUDA, Bash Scripting, Cloud Computing
Professional Experience
Data Scientist, Unconventional Reservoir
WD Von Gonten Laboratories | Houston, Texas, USA | June 2018 – now
Fast 3D geomechanical cube model (Fully written in Python) to generate structural and geomechanical properties
- Database querying (SQL), EDW (Snowflake-connector), Relational Database Management (Access, SQLite)
- Geographical data manipulation (projection, processing and mapping), using python’s pyshp, GeoPandas, pyproj
- Predictive modeling| structural modeling, kernel hyper-parameter-optimization, implementing various interpolation, geostatistical techniques, Generalized Linear Models, GP Regression (scikit-learn, GPFlow using Tensorflow)
- Software development version control (BitBucket, GitHub, git), UI sketching, handling google maps and satellite images, implementing various interpolation, geostatistical techniques
- Calling R from python, using various R packages (in POC, replaced later with python and other more efficient packages)
Deep-learning convolutional models to extract information from logs and images
- CNN using Keras| train network using FMI and pseudo-density log images to mimic CT images
- Image inpainting techniques to complete FMI | interpolation, multiple-point geostatistics, pattern filling using DL
- Depth registration to align core-depth, and measured-depth
- POCs: Well2Well auto-correlation, DL for automatic labeling sedimentological features, Fracture density prediction
Python-based reservoir visualizer using Mayavi, and Paraview
- Computer graphic| VTK , TVTK, python scripting to generate vtk file, converting RESQML file to VTS (structured grid)
- Combining data, generating trans-sectional view, filtering and processing of graphical outputs
Shale Reservoirs 3D Modeling
- Frack Modeling Project for Nobel Energy, BPX, Vista Energy, Oasis Petroleum
- Visualizing clients’ pad-data| Matplotlib, Seaborn, ArcGIS, poltly, Data Studio, QGIS
- Working with client data, exported python’s 3D cube results, comparing and testing in Petrel
- Revisiting models using operational data|FMEA procedure
- Predicting Fracking Performance using completion, drilling and geological data in vaca-muerta formation
Research Associate, Geostatistical Reservoir Modeling
University of Wyoming | Laramie, Wyoming, USA | Oct. 2015–Sep. 2017
Improved Geostatistical Reservoir Modeling methods
- New framework that generated fine-scale geological models of meandering reservoirs guided by seismic data
- Statistical learning| hierarchical pattern similarity technique and K-mean clustering
- MATLAB scripting, synthetic seismic data generation, Forward modeling, wavelet deconvolution, pattern distance calculation)
- Published in Mathematical Geosciences
Assessed the impact of geological uncertainty on fluvial reservoir evaluation
- Using process-based fluvial reservoir models, Investigating the impact of geometrical parameters of channels and avulsion periodicity on models’ NGR and cumulative production
- Facies-based petrophysical property simulation, reservoir modeling using Petrel, Eclipse and StreamSim
- Published in Petroleum Sciences
Researcher, Subsurface modeling of aquifer systems
University of Delaware | Newark, Delaware, USA | Oct 2013–Sep 2015
Assessed aquifer vulnerability to arsenic contaminants in the Bengal Delta and determined that the threat of contamination of deep aquifer resources due to over-pumping in the region has been underestimated.
- Data augmentation, +50 different connectivity metrics to predict size and center of saltwater mixing zone
- Using Lasso and Ridge regression, Geologic influence on groundwater salinity drives large seawater circulation through the continental shelf, published in Geophysical Research Letter, 2016.
- Published in several high-ranking journals, e.g. Nature Communication
Simulation lava-flow emplacement that can be used to evaluate lava's inundation zone
- Monte-Carlo simulation of flow path and inundation probability, extracting geological feature of mapped lava flow (ArcGIS)
- A physic-based and hybrid model proposed to generate aquifer scale heterogeneity of volcanic deposits, MATALB and C++
Performed QA/QC of a database of Bengal Delta lithological data
- Data scrubbing, statistical and spatial connectivity/continuity of lithological data, subsurface mapping (Access , VBA)
- Using Arc GIS and Rockworks to map and combine Geological, and well databases. Python scripting in SgeMS and arcpy
Assistant Professor & Lecturer
Shahrood University of Technology | Tehran, Tehran, Iran | May 2008–Sep 2013
Member of the Research Committee, Member of the Scientific and Organizing Committee, Academic Advisor, Guest Lecturer at University of Tehran and Tarbiat Modares University
Taught graduate and undergraduate courses
- Geostatistics, Statistics and probability for mine engineers and geoscientists (mining, petroleum), Principle of mineral exploration techniques, including remote sensing geomatics, field geology, exploration and mining geology, principle of resource/reserve estimation, mine economy risk evaluation, non-metallic deposit exploration, geological simulations, industrial minerals and rocks (geology, characterization, and application)
Consulting and Research
- Conducted research for industrial partners; performed characterization and modeling of Esfordi phosphate ore classes; wrote several research proposals
- Organized workshops, including Statistical Methods and Experimental Design for Mineral Processing and Mineral Economy and Marketing
- Resource-reserve estimation of Narigan deposit, Sari-Gunay gold deposit and Esphordi Phosphate deposit
- Spatial factor analysis (spatial filtering), Comparison of factorial kriging analysis method and upward continuation filter to recognize subsurface structures—a case study: gravity data from a hydrocarbon field
Education
PhD in Mining Geostatistics | MINES ParisTech | Paris, Paris, France| 2003–2007
Centre de Géosciences, Ecole Nationale Supérieur des Mines de Paris (ParisTech)
Master of Engineering in Geostatistics | MINES ParisTech | Fontainebleau, Seine-et-Marne, France| 2005–2006
CFSG, Centre de Géostatistique, Ecole Nationale Supérieur des Mines de Paris (ParisTech)
M.Sc. in Mining and Mineral Engineering | University of Tehran | Tehran, Tehran, Iran | 1997–2000
Specialized in Mineral Exploration Engineering
B.Sc. in Mining Engineering | University of Tehran | Tehran, Tehran, Iran | 1993–1997
Specialized in Mineral Exploration Engineering
Qualification
Courses & Training: Machine Learning, Mineral Potential Mapping, ArcGIS, Mineral Resources Development (CLAIM), GIS and RS, Mining Engineering training in Pb-Zn mine (Anguran, Zanjan)
Other languages: French