Data Python
Resume:
Stephen Pohl
Weatherford, TX 76087
*******@*****.***
EDUCATION
University of Texas at Dallas, Richardson, TX —B.S. Data Science 3.0 GPA
Fall 2016 – Fall 2020
Weatherford College, Weatherford, TX
Fall 2015 – Spring 2016
EXPERIENCE
University of Texas at Dallas, Richardson, TX – Student Researcher January 2020 – May 2020
● Reformatted dirty data into cleaned “ready to read” data
● Stemmed and took out stop words from data for vectorization prepping
● Created two vectorized data sets, one using Word2Vec and the other using Count Vectorization
● Applied SVM, MLP learning methods onto the vectorized data resulting In an above 95% accuracy
SignaPay – Optical Character Recognition (OCR) Capstone January 2020 – May 2020
● Used the Keras functional API in TensorFlow to create a model following a ResNet model structure
● Wrote and implemented custom CTC loss function for the model
● Created custom metrices to calculate character and word error rates as well as average distances of predicted against true values
● Model predicted handwritten characters with above 95% accuracy and Handwritten words with above 85% accuracy
CIFAR10 CNN Identification Project Using TensorFlow August 2019 – December 2019
● Created a CNN to identify CIFAR10 images with above 60% accuracy
● Attached a Pretrained CNN to further increase the image identification accuracy to above 90% accuracy
● Wrote IEEE paper exploring the advantages of adding a pretrained network on an existing CNN
MNIST CNN Identification Project Using PyTorch
August 2019 – December 2019
● Randomly sampled 100 of each digit from MNIST data set to create random train and test splits
● Created simple model involving convolutional layers followed by a softmax using PyTorch
● Tested the model using AdaBoosting with decision trees and logistic regression resulting in an above 80% accuracy
SKILLS
Python
Java
C++
Some Experience with C
Some Experience with R
Some Experience with SQL
TensorFlow 2
Keras API
Scikit-Learn
Pandas
Numpy
Object Oriented Design
JavaFX
Clean Dirty Data using Python
Some Experience with PyTorch
Some Experience with Java Swing
Google Colab
Eclipse IDE
RStudio
CodeBlocks IDE
Some Experience with Anaconda
Contact this candidate