Welcome!


Welcome to my website, my name is Steve Bottos. I am currently working towards a research-based Master's degree (MaSc) in Electrical and Computer Engineering, with my research dealing mainly with the realms of Machine Learning and Data Science. Here you'll find some information about myself as well as some past projects and works-in-progress. I hope that you enjoy my work as much as I enjoyed producing it. For more about me, head to the "About" section. Check out the feed below for some of my most recent work, or check out the "Projects" page.

Gaze Tracking System

Here you’ll find a Jupyter Notebooks writeup detailing Version 1 of the Gaze Tracking System. The Gaze Tracking System designed to determine which line of text that an individual is currently reading given absolutely no prior knowledge of the page layout or what is being displayed on the screen whatsoever. Currently, the algorihm is able to detect when a reader progresses from one line to the next down the page, beginning from the top, and is able to accurately pinpoint in real time when this change of lines takes place. Given noisy data, the system performs well, and is able to accurately construct the layout of the page in terms of lines of text on the page in real time.

If you’re interested in just the code, head here.

Key Concepts: Machine Learning, Python (Numpy, Pomegranate), Bayesian Statistics, Hidden Markov Models

Cognitive Context Detection With Matlab

This repository contains a program which employs a Neural Network (NN), written by me from scratch, designed to recieve input data in the form of an individual’s pupil diameter in order to detect whether or not the particular individual is currently engaged in a task which they consider to be easy, somewhat difficult, or very difficult. A short training/calibration period is necessary (approximately 20 minutes - the duration can be adjusted depending on necessity), during which time the NN learns the specific parameters of that individual. After the training period the program is designed to take measurements every 2 minutes using data from that time frame to construct a single feature vector, outputting the level of focus that the individual was most likely experiencing during those two minutes.

Key Concepts: Machine Learning, Neural Networks, Matlab, Feature Engineering

The Hockey Summit Website

The Hockey Summit Website

TheHockeySummit.com is an ongoing project of mine. The coach of the San Diego Gulls and founder of THS, Justin Roethlingshoefer, had approached me asking if I was available to work on the site for him in early 2018. I was, and still am, a bit of a web-dev hobbyist so I welcomed the opportunity to work on something new. I currently manage any front-end or back-end changes to the site that must be implemented as the company grows.