Justin Essert

Machine Learning Engineer

OBJECTIVE

I am a lead machine learning engineer with a focus in time series modeling.
I place a heavy emphasis on building models and software in a reusable and modular way so that the work can be built upon and leveraged by multiple use cases. Therefore, I aim to build up a suite of sustainable tools for time series modeling that can be easily leveraged and expanded.
People have described me as an effective communicator with the ability to lead and influence people towards a shared goal. I have frequently driven collaborations between different teams so that we can work together on a shared solution saving us both time and resources. Additionally I am consistently tapped as a mentor to younger developers to help them grow their modeling and software engineering skills.
I am looking for opportunities to remain in a technical role while also driving use-cases in a tech-lead capacity in the space of machine learning engineering.

LANGUAGES & TECHNOLOGIES

Languages

  • Python
    • Pandas/Numpy
    • Keras
    • Tensorflow
  • Typescript
    • Angular
  • Javascript
    • AngularJS
    • NodeJS/ExpressJS
  • C
  • Java
  • MongoDB
  • SQL / mySQL

Technologies

  • Time Series Modeling
    • Various Forecasting Models
    • Various Anomaly Detectoin Models
    • Various Change Point Detectoin Models
  • Deep Learning
    • Generative Adversarial Networks (GANs)
    • Convolutional Neural Networks (CNNs)
    • Long Short-Term Memory (LSTM) Networks
  • GIT Version Control
  • Agile Project Management
    • Scrum Practices

WORK EXPERIENCE

CAPITAL ONE - CENTER FOR MACHINE LEARNING

Principal/Lead Machine Learning Engineer | Capital One Data Insights | Jan 2021 - present
  • Built general models for detecting anomalies and changepoints that can be used across many use cases / teams.
  • Prevent billions of dollars per year of fraud by detecting anomalies in transaction data.
  • Provide easy-to-use interfaces for users to plug into machine learning models.
Machine Learning Engineer | ML Consulting with Credit Loss Forecasting | Sep 2018 - Jan 2021
  • Built models to forecast credit loss that outperformed existing well-established models.
  • Built orchestration framework to run generic machine learning models easily for business users.
Machine Learning & Data Engineering Intern | Summer 2018
  • Provided in-house machine learning consulting to teams across Capital One in order to improve efficiency and deliver better business insights.
  • Researched & implemented methods at the forefront of machine learning innovation such as Generative Adversarial Networks (GANs) and Long Short-Term Memory (LSTM) networks.

DIVISION OF INFORMATION TECHNOLOGY - UW MADISON

Senior Consultant | January 2017 - Present
  • I lead a team of 3-6 members in assisting dozens of customers with technology sales and accurate technical support.
  • Designed & lead the development of an appointment application for the DoIT Walk-In Help Desk
Tech Store & Walk-In Help Desk Consultant | August 2015 - December 2016
  • Learned to effectively and clearly communicate, verbally and in writing, with both technical and non-technical people.
  • Efficiently and effectively diagnosed and repaired problems with operating systems, applications and hardware defects.

UNITED HEALTHCARE DIGITAL - DIGITAL DATA ANALYTICS

Data Analyist & Project Management | June 2017 - August 2017
  • Used tools such as Python & Adobe Analytics for data manipulation and analysis to provide financial insights.
  • Helped develop a web app providing data analysis & visualization, allowing management to make data-driven decisions.
  • Explored my passion of machine learning and how it exists in a business setting.

DUNKIN' DONUTS

Administrative Assistant | January 2015 - August 2015
Shift Manager | March 2014 - August 2015
Crew Member | August 2012 - February 2014
  • Developed strong relationships with my customers and my team with outstanding customer service and communication.
  • Managed a team in a fast-paced work environment with the ability to think quickly and successfully handle setbacks.

EDUCATION

UNIVERSITY OF WISCONSIN MADISON - CLASS OF 2019

M.S. Machine Learning & Signal Processing
  • Pursued a deeper and more theoretical understanding of machine learning through a one year masters.

UNIVERSITY OF WISCONSIN MADISON - CLASS OF 2018

B.S. Computer Sciences & Electrical Engineering (Double Major)
  • Studied computer science & electrical engineering with a focus in machine learning.
  • Selected classes that involve pattern recognition, machine learning, and optimization.
Physics Certificate
  • Completed a physics certificate to better understand the nature of Quantum Mechanics and to explore how to create efficient algorithms for a quantum computer.

MIDDLETON HIGH SCHOOL - MIDDLETON, WI

Class of 2014
  • Captain of the Middleton Debate Team
  • Captain of the Middleton Forensics Team
  • Member of the Middleton Varsity Mock Trial Team
  • Section Lead in the Middleton Jazz Band

Personal Projects

MARCH MADNESS MASCOT BRACKET

Javascript, HTML | Mascot Bracket Constructor
  • Each year my sister and I have created a March Madness bracket basing our decisions on which mascot we like better.
  • I created a static website to make the process of creating this bracket easier.

DISC GOLF STATISTICS

Python, Seaborn, & Matplotlib | GitHub: hierarchical-deep-cnn
  • Combined interests of data analysis and disc golf by creating a product that generates a report based on your disc golf stats.
  • Generates Markdown reports containing current stats (example report)

HIERARCHICAL DEEP CONVOLUTIONAL NEURAL NETWORK

Python, Keras, & Tensorflow | GitHub: hierarchical-deep-cnn
  • Developed a model that based on a 2015 study by Zhicheng Yan et al. by using a hierarchically structured CNN to analyze and categorize the images within the Cifar100 dataset.

FOOTBALL PREDICTION ALGORITM

Java & HTML
  • Engineered an application that analyzes ESPN data and then predicts the outcome of NFL games using machine learning.
  • Produced a week-to-week accuracy of 67-75%.

NEURAL NETWORK RESEARCH

CONTACT ME

Feel free to reach out and say hi!

Email: justinkessert@gmail.com

Copyright © Justin Essert 2017