I'd love to hear from you.

  • github
  • Black LinkedIn Icon
  • Black Twitter Icon
  • Spotify Social Icon

HELLO, I'M

Rob Firstman

Computer Science Student

Rob Firstman

Undergraduate in computer science concentrating in artificial intelligence and computer networking

Atlanta, Georgia

  • github
  • White LinkedIn Icon
  • Black Twitter Icon
  • Spotify Social Icon
 

About

MY BACKGROUND

I am a fourth year undergraduate computer science student at Georgia Tech with concentrations in artificial intelligence and computer networking. In my free time, I enjoy teaching myself new coding skills and applying those to side projects. With enterprise development experience and a passion for technology and learning, I am prepared to succeed in the development world.

 

Education

WHAT I’VE LEARNED

2016 – Present

Georgia Institute of Technology

Degree: B.S. in computer science

GPA: 3.80/4.00

Concentration: artificial intelligence and computer networking

Expected Graduation: May 2020

CS Courses: machine learning, systems & networking, computer organization & programming, computer networking, user interface design, information visualization, data structures & algorithms, design & analysis of Algorithms, relational databases

Experience

WHERE I’VE WORKED

May – August 2019

Microsoft- Software Engineer Intern

  • Developed an extension to Windows Admin Center

  • Enhanced the setup process for Azure File Sync

  • Contributed to and fixed bugs for production code

  • Utilized online developer tools to track progress and provide feedback

Nov. 2017 – Present

The Home Depot - Software Engineer Intern

Description: Full-time during Summer 2018, part-time during the school year.

  • Developed and deployed two full-stack web applications

  • Collaborated with rounded teams to tackle complex and changing requirements

  • Worked alongside user experience designers and business managers to maximize usability and value

Skills & Languages

WHAT I BRING TO THE TABLE

Web Technologies

React.js

Node.js

Angular

Redux

Express

d3.js

Programming Languages

Python

JavaScript

Java

Go

C

SQL

Other Technologies

TensorFlow

PyTorch

Scikit-learn

Docker

 
 

Projects

PROFESSIONAL AND PERSONAL

MyFloorView

Web-based data viz tool created for The Home Depot

  • Performed user and business research to determine the project's direction

  • Provided a means to improve warehouse efficiency while saving more than $10k over costlier proposed solutions

  • Made with React.js, Go, and Oracle SQL DB

Coinrun Deep Q-Learning

A deep learning approach to reinforcement learning

  • Utilized PyTorch to train an agent to play a platforming game

  • Created a convolutional neural network trained on game footage

  • Experimented with various network designs to maximize performance

  • Produced a model that could routinely beat testing levels

Robo Recruit

Recruitment utility created for The Home Depot

  • Developed to streamline the recruitment process

  • Implemented a parser to extract important info from resumes

  • Enabled recruiters to spend less time on data entry and more on the recruits themselves

  • Made with React.js, Go, and PostgreSQL

Hip Hop/Jazz Analysis

Tools to assist in getting and analyzing rap lyric data

  • Created a tool to scrape lyric data using the Genius API

  • Organized this data into individual verses sorted by artists

  • Generated visual representations of rhyme schemes

  • Utilized regular expressions and phonetic encodings to produce these tools

 

Interests

One of my favorites things in this world is listening to and analyzing music, whether it be alternative rock, rap, soul, funk, or jazz. I grew up in a family that loves to listen to and play music, so I grew up learning how to play the drums and later the guitar. I enjoy any opportunity to discuss music regardless of genre. I also enjoy following the NBA, and I am an avid fan of the Atlanta Hawks. In the past, I have attempted to utilize the wealth of statistical data found in the NBA to train machine learning models. One such model was predicting a player's position based on their box score stats. The model was ultimately unable to reach accuracy over 60%, but it was a good thought experiment. In my free time, I enjoy playing video games or going out into nature to get a break from all the screens.