Data Machine

SKILLS

PROGRAMMING

Proficient in python and popular libraries:

• Pandas

• NumPy & SciPy

• StatsModels

• Matplotlib

• Scikit-learn

• XGBoost

• PyTorch

Experiences with C++ / JAVA / MATLAB / C#

Familiar with MySQL, parallelism

MACHINE LEARNING

Regression and classification

• Support Vector Machines

• Neural Networks

• Naïve Bayes

• K-Nearest Neighbors

• Random Forest

Cluster Analysis

• Gaussian Mixture Model

• Hierarchical Clustering

• K-Means

• DBSCAN

Time series analysis

Dimensionality Reduction

Familiar with Natural Language Processing

EDUCATION

MS, COMPUTER SIENCE

Virginia Tech, GPA: 4.00/4.00

Augu 2019 – present Blacksburg, VA

Thesis: Application of Machine Learning in

Geo-energy Exploration

Ph.D., PETROLEUM ENGINEERING

Texas Tech University, GPA: 4.00/4.00

Aug 2015 – Aug 2019 Lubbock, TX

Thesis: Software & Modeling Development

for Wax Deposition Phenomenon

LINKS

linkedin.com/in/arya-shahdi

https://github.com/aryashahdi

EXPERIENCE

Virginia Tech Research Assistant

Aug 2019 – present Blacksburg, VA

• Developed a state-of-the-art ensemble clustering-regression model to group data into clusters using GMM to mitigate the class confusion problem and then use ANN for class label prediction. validation.

• Applied deep learning, rigid regression and random forest for subsurface temperature prediction using geological setting.

• Currently performing time-series analysis for hydrocarbon production rate and crude oil price estimation using classic time series models in addition to LSTM Recurrent Neural Networks. Texas Tech University Ph.D. Candidate & Researcher Aug 2015– Aug 2019 Lubbock, TX

• Developed two software packages for oil & gas industry containing 21,000 lines of code in C++ and Python coupled with a C# GUI. OpenMp parallelism techniques have been applied to enhance the performance (by nearly 5 times). Advanced numerical methods were programmed (e.g., Dorman Prince, Levenberg-Marquardt optimization algorithm, etc.).

• Applied Self-Organizing Map (SOM) for clustering and developed sets of Generic-Algorithm-Based Correlations for CO2 solubility estimation in deep aquifers for sequestration applications.

• Developed a simulation tool to read through real-time capacitance censor data in pipe and fit time-series models to predict various characteristics of the flow regime.

Azad University Machine Learning Researcher

Set 2012 – Aug 2014 Tehran, Iran

• Collaborated in a research project to estimate compressibility factor in gas systems. A LS-SVM model was developed which resulted in higher accuracy in compare to all other physics-based methods.

• The developed model in the previous study was then used to predict Frictional Pressure Loss in Inclined Annuli in a separate research effort and resulted in a state-of-the-art performance (R2 = 0.95, ARD = -1.68%)

RELEVANT COURSES AND CERTIFICATIONS

University courses: Data Analytics (CS5525) - Advanced Machine Learning (CS5824) - Data Structure and Algorithms (CS3114) Coursera: Practical Time Series Analysis - Sequence, Time Series and Prediction - Java Programming: Arrays, Lists, and Structured Data - Solving Problems with Software

DataCamp: Time Series Analysis in Python - Deep Neural Networks with PyTorch – Pandas Foundations - Feature Engineering for Machine Learning in Python

PUBLICATIONS

Published and presented five papers in peer-reviewed journals and conference proceedings

Arya Shahdi

US Permanent resident

703-***-****

**********@**.***

Contact this candidate