I'm Harlan Ferguson, a full-stack software developer. I graduated with an Advanced Diploma in Computer
Programming and Analysis at George Brown College, Toronto.
I enjoy programming projects primarily in JavaScript and Python but I also have a large knowledge of many
different languages and tech stacks. I'm eager to start working in the industry.
My journey in tech is marked by a diverse portfolio of projects—from creating a social media site,
to leveraging AI for enhancing email security. I've worked in industries that constantly require problem solving
and I'm passionate about learning and improving my skillset.
My ability to rapidly adapt and deliver under tight deadlines doesn’t just reflect my technical
proficiency but also underscores my passion for making a tangible impact through code.
As I navigate the beginnings of my career, I'm eager to collaborate with forward-thinking teams,
contribute to meaningful projects, and grow into a role where I can blend technical innovation
with strategic insight. Below, you’ll find some of my recent projects that showcase not just
what I’ve learned, but what I love doing.
Projects
Emotion Detection with EfficientNet
This Emotion Detection Project showcases the development and refinements of a model
aimed at identifying human emotions. Initial overfitting was tackled by introducing
validation sets and early stopping, without much success. Progressively, through
adjustments such as data augmentation, model simplification, and convolutional layer
enhancements, slight improvements were achieved. Transitioning from TensorFlow to PyTorch,
the project saw significant enhancement with the adoption of EfficientNet, ultimately
achieving 68.5% accuracy, which is impressive considering the "high score" for this dataset sits around 75%.
This journey not only refined my understanding of neural network optimization but also enhanced my
skills in leveraging GPU over CPU, optimizing performance significantly.
My-Game, our capstone project at George Brown College, is a comprehensive sports
social media platform that simplifies team management and enhances user interaction.
As the project manager and lead developer, I spearheaded its creation, focusing
on robust features like Auth0 integration for secure login, a dynamic feed for
user posts, and real-time messaging with Socket.io. Utilizing Firebase for its
CDN, the app delivers media content swiftly, ensuring an engaging user experience.
Comprehensive user profiles, direct messaging, friendship, and post CRUD are among the
key features I implemented, alongside secure JWT authentication for protected routes.
My-Game exemplifies a blend of technical excellence and user-centric design,
showcasing my commitment to developing scalable and secure web applications.
For a dive into Angular and real-world API consumption, I built a SpaceX Mission Tracker.
This app hits the SpaceX API to list missions, filter them by year, and dive into mission
specifics, including embedded launch videos for a richer experience. Despite being a
first-timer with Angular, I managed to get a grasp on its ecosystem, from routing to services,
and how to properly architect an Angular application.
Facing challenges like hosting issues and making sure Angular Material played nice with custom
styles, I leaned into problem-solving and debugging - crucial skills in any developer's toolkit.
The outcome: a functional, informative app that reinforces my fundamentals in Angular and my commitment
to learning by doing.
Developed within a tight 12-hour deadline for a school lab test, this chat application
leverages the Dockerized MERN stack for a complete user registration and login
experience, using JWT for secure authentication. Utilizing Socket.io, it enables
real-time communication across various chatrooms, with the server dynamically
notifying users upon new arrivals. Each chatroom maintains its unique chat history,
ensuring data persistence and continuity. User information is stored securely in
MongoDB, adhering to best security practices.
This project not only solidified
my understanding of websockets and live communications but also demonstrated my
ability to rapidly develop a feature-rich application under time constraints.
In my role as the project lead for "City Scav", I guided a talented
team in designing and implementing an interactive and user-friendly mobile application. Developed using
Android Studio, Java, and Gradle, this app integrates the Google Maps API, allowing
users to engage in a digital scavenger hunt. The app features custom markers for
points of interest (POIs) which users collect and track on a dedicated screen.
We incorporated an achievement system to reward users based on the number of collected
POIs and included social sharing capabilities for enhanced user interaction. The app's
design reflects an outdoorsy aesthetic with AI-generated colors and images, aimed at
providing an enjoyable user experience.
This project not only showcases my technical skills in mobile app development but
also highlights my capabilities in project management and creative design. We recieved a 100% grade
from our professor for this assignment.
Achieving a top grade of 100% in my fullstack class's final assignment, I developed an Employee
Management Web App using the MERN stack.
This web application provides a secure login system with robust password encryption,
leading to a dashboard that displays key employee metrics. The dashboard features a
pie chart showing the gender distribution, average salary, and total employee count,
offering managers immediate and valuable insights.
A critical feature of this app is the ability to perform CRUD operations on employee
data, enabling effective team management. The app's frontend, designed with React.js
, offers a user-friendly experience, while the backend, using Node.js and Express.js
with MongoDB, ensures reliable data management and server-side operations. This
project demonstrates my capability to blend technical proficiency with practical
business solutions in a full-stack development environment.
Chrome Extension that blocks phishing links using AI
Developed in a 48-hour hackathon, this Chrome extension leverages AI to enhance email security without
any user setup. Once installed, it seamlessly integrates with your Gmail. The extension scrutinizes
email senders against a pre-defined whitelist. Emails from trusted domains receive a green dot,
indicating safety. For others, it employs a Random Forest classifier, trained on phishing email
datasets, to assess potential risks. Detected phishing attempts are marked with a red 'phishing'
alert, and hyperlinks are disabled to prevent access to potentially harmful sites. Safe emails are
similarly marked with a green dot, reassuring the user of their legitimacy. This solution blends
simplicity with robust security, offering a user-friendly approach to email protection.
For this project, I wanted to make my own dataset and then train a language model. I went with Reddit because of their free API use. After
generating a dataframe with thousands of subreddit titles and their corresponding subreddits as their labels, I cleaned and vectorized
it in preparation for training. Using Scikit's Bayes model, I trained and tuned the data to an average F1 score of 80%. From there,
I developed a webapp for users to enter a title, and have ML predict where that title would belong on Reddit.
For this project, we were tasked to make a management system. I chose to manage repair workorders
because at the time of writing this, I'm working part-time as a maintenance worker at a building.
It allows for basic workorder creation, search, editting, and deletion all with a simple UI.
For this project, my team and I developed a Tic-Tac-Toe game using Java. For the GUI, we used Swing to get a better understanding of that package.
The menu allows for the user to choose between a human player, a computer player using random inputs, or a computer player using the minimax algorithm.
It was our introduction to "AI" and recursive algorithms since we had to write the algorithm ourselves and not use imports.
For this project, I was tasked to create a console application for a store that can manage employee purchases.
Part of the project requirements were to give different discounts to certain employement types, and to track total purchases versus total discounts.
I went above and beyond by using best object oriented programming practices, which was not a requirement, and not taught in the course
for another few weeks. Going this design route allowed for easier code readability and understanding, compared to how others implemented
the assignment, by writing all of it in a single python file.
This was my first ever project (2020). It originally used SMTP to send birthday messages to my friends. It's what gave me the confidence to go to school
full-time for programming and I'm glad I did.
It was fun to learn/write at the time, but it had it's flaws. Like how most mobile carriers don't support SMTP by default. So it ended up not being as helpful
as planned. So I figured I would re-write it from scratch using my current expertise. Now the program uses the Twilio API, which allows me to text anyone in my
dataset without issue. Now the program is a lot more reliable. Successfully running with cron jobs on a Raspberry Pi3!
