Hello, World!

Welcome to the homepage of
Zachary McCoy
I am a software systems enthusiast

About Me

I am a Machine Learning Engineer with 10+ years of professional software development experience, and an expertise in Integrations. It has been said that 80%+ of an ML project is data preparation, and it is a particular skill of mine to be able to find, organize, analyze, and augment datasets. I've assumed roles ranging from Client Liaison to DevOps to QA (and several in-between), and I've successfully collaborated with colleagues and teams across cultures, time-zones, and companies.

University of Pittsburgh

B.S., Computer Science - Apr 2005
   (With High Honors; GPA 3.6)

While I had been programming computers since I was young, the University of Pittsburgh presented an opportunity to hone my skills and be challenged to learn technologies that I might otherwise have ignored. During my 4 years at Pitt, I spent time around some very smart people who were innovating in the fields of computer science, neuroscience, education, and medicine. It was first experience where programming was overlapping with other disciplines, and where software was having a very tangible impact on the lives of people. Early on in my undergraduate studies, I took a role of programming GUIs to "wrap" command-line based fMRI programs.

Udacity | MLND

Machine Learning Nanodegree - Apr 2018

  • SmartCab (Q-Learning) - Applied Reinforcement Learning to build a simulated vehicle navigation agent

  • Customer Segments - Unsupervised Learning with PCA to discover patterns and natural categories

    • Coursera | DeepLearning.ai

    DeepLearning.ai Specialization - Jan 2018

  • Neural Networks and Deep Learning
  • Improving Deep Neural Networks:
    Hyperparameter tuning, Regularization and Optimization
  • Structuring Machine Learning Projects
  • Convolutional Neural Networks
  • Sequence Models

    • Interesting project work



    NIH Chest X-Ray Classifier

    Inspired by Stanford ML Group's CheXNet, I also decided to train a DenseNet network to perform binary classification on the same NIH chest x-ray dataset, albeit on a single category ("Pulmonary Fibrosis"). In this case, the binary classification is a one-vs-all (a.k.a. one-vs-rest) approach for the lung disease category: "Pulmonary Fibrosis"; samples from this single class are taken against a random sample of the other 14 categories (including "No Finding").

    Sentiment Analysis (Arabic)

    A deep learning (LSTM) sentiment analysis project to determine positive/negative sentiment in Arabic social media content. This project makes use of Arabic word embeddings (https://github.com/iamaziz/ar-embeddings) and a dataset of 2000 Twitter posts (https://archive.ics.uci.edu/ml/datasets/Twitter+Data+set+for+Arabic+Sentiment+Analysis)

    One-Shot Audio (voice recognition)

    Experiment with "one-shot learning" techniques to recognize my voice signature. This was my submission and Live Demo for LVTech hack-a-thon 2018. I won the prize for "Best Solo 'Team'"