--Introduction
Hello.
My friends think I'm a polymath, but I think that I'm just me.
I have skills in software, game development, data engineering, and machine learning.
I like putting them all together to do cool things.
I am a multi-instrumentalist, songwriter, and gamer.
I am a cancer survivor and a martial artist.
Legends have been told that I even whisper to dogs...
Project 1
Starcraft 2 Reinforcement Learning Agent
Using Deepmind's PySC2 library, I created a Zerg AI Agent that uses reinforcement learning with a Q-Learning table to play the game. The project has a few different AI's that were built such as an incredibly basic bot as a starting point for anyone else who wants to expand on the concept. I also explored creating a Behavior Tree that included the necessary node types such as Control, Action and Conditional nodes. However, SC2 has HUNDREDs of actions available and a heuristic agent would not be able to compare the prowess of the RL Agent. This project is ongoing, but it demonstrates my natural ability for game development and machine learning, my ability to quickly jump into a new library, and the ability to improve on project structure by using automation tools such as Doxygen, Logging and even ChatGPT for creating comments quickly, allowing for an expedited documentation process. Future hopes for this project is extracting the used feature data to send through a data pipeline, store it properly, and provide dashboards and analysis for the training features.
Project 2
NASA Rover ETL Pipeline
I combined my passion to explore data and engineering techniques in this Databricks ETL pipeline. The pipeline connects to the NASA Mars Rover API and extracts Mars Rover camera and photo data. I utilized a medallion architecture along with storing the data in AWS S3 locations forming a data lake. After cleaning and transforming the data, it is used to create various dashboards. One sample dashboard displays camera, mission and rover data. Another dashboard does image analysis on a photo extracted from the data lake. Technologies used were Databricks, Python, Pyspark, SQL, and AWS.
Project 3
BabyWatch Baby Monitor
I led a team of 3 engineers to create an AI-powered encrypted baby monitor. The project was deployed on a Jetson Nano using Python, Flask, OpenCV, and Sci-kit Learn. I wrote the Flask App, and the firmware for both the server and the camera on the Jetson Nano. I also built a custom data-set using a baby doll and trained various models with it. I pickled the best model and deployed it to the Jetson Nano, where it ran its machine learning algorithms in real time to notify a user if their child was in danger. The machine learning repo can be viewed here. The code for the server and camera to be deployed on a Jetson Nano is here.
Project 4
YelpCamp
Yelpcamp is a campground review site built with Node.js, Express, and MongoDB. The full stack application features user authentication (sign up, login, logout), user profiles and campground ownership, campground ratings and reviews, and map integration. It also supports CRUD operations for campgrounds. This project is based off of Colt Steele's Web Developer 2024 bootcamp. You can check it out here: YelpCamp. Note that it might take up to a minute for the cluster to start if you're the first visitor in some time. This project is very old and has a few bugs due to updates.