Long_USD-AAI_Portfolio
Johnathan Long
AI Engineer / Scientist
Welcome
I’m Johnathan and I will graduate this summer with a Master’s in Applied Artificial Intelligence from the University of San Diego. I also have a Bachelor’s in Physics, so I really enjoy grasping theoretical concepts and being able to find impactful, almost immediate, use-cases for them with artificial intelligence. My prior experience includes material-science & engineering, and education. I pride myself in being a multifaceted technologist with a strong science background, and am open to opportunities of growth and development. Currently, continued education and research specializing in computer vision to detect neurodegenerative diseases sounds like an attractive option.
Q2-Resume(25)
In the Works
(1) USD Tech Challenge Competition (2025)
For my university’s annual Tech Competition, I am submitting a AI-enabled Micro Aerial Vehicle intended for personal accomodation. A sleek energy conservative design allows for maximum battery life while computer vision algorithms are used for recognition and proximity. My background in Physics supports me in exploring the depths of fluid mechanics when considering a unique aerial design based upon Bernoulli’s Principle. My skills in artificial intelligence presents me with the skills to build functional algorithms to automate the drone. All code used for the project will be uploaded here after the project deadline on April 15th, 2025.
*Project is at concept stage and building CAD model for CFD simulation
_(The following is NOT the functioning model but only a conceptual visual designed in blender)
(2) Teaching with ChatGPT - Workshop (2025)
Here I host a workshop for teaching with generative AI. The workshop covers the basics of using generative AI, the limitations of AI, what to expect of academic policies in the near future, and so on. This workshop was limited to USD’s School of Business which exhibits my ability to interpret highly technical and relevant content into communications digestible through various sectors. These skills are transferable to technical writing, project managme
nt, and Science Leadership/Managment. See brief sample of my curriculum design here.
Completed Projects
(3) Heart Disease Prediction Model (2024)
This project explores the fundamentals of Artificial Intelligence (AI) by comparing two machine learning algorithms, Random Forest and XGBoost, for predictive analysis using the Heart Disease Prediction dataset. Exploratory Data Analysis (EDA) reveals key trends, such as a strong correlations and highlights imbalances that may introduce bias. This project emphasizes machine learning development and real-world applications.
Files:
Heart_Disease_Prediction.csv = Dataset
Long_AI_Fundamentals.ipynb = Notebook
Long_AI_Fundamentals_Notebook.pdf = pdf version
(4) Object Detection - Computer Vision Exercise (2024)
Introducing key computer vision methodologies through a three-part assignment involving video processing, pose estimation, and real-time object detection. OpenCV is used to read a video, extract metadata, and apply filters to create a sketch effect. Pose estimation using TensorFlow Hub to analyze movements in a GIF format video. Real-time object detection with YOLOv8 and a webcam integrates pre-trained models for live camera feed capture. Note you must provide your own video files (video/path) to use the code effectively. This project mphasizes hands-on learning with tools like Google Colab, OpenCV, and TensorFlow, progressively building skills in video analysis and AI-based detection techniques.
Files:
YOLO for Object Detection.ipynb = Stand Alone Notebook
(5) Building a Chatbot (2024)
In this study we build a chatbot by leveraging the Ubuntu Dialogue Corpus, a vast dataset containing nearly one million dialogues from Ubuntu support chats. We train a chatbot on a focused subset of dialogueText_196 we used for processing. We used a pretrained transformer-based GPT architecture. Once trained, the chatbot was integrated into a user-friendly interface. Responses are displayed within the notebook for easy tests and continuing research. The project demonstrates how naturally occurring dialogue data can enhance a chatbot’s ability to engage users effectively.
Files:
Chatbot_Ubuntu.ipynb = Notebook (Run the code from the checkpoint to access the chatbot)
checkpoint-500 = Model (Load this model checkpoint to prevent retraining)
___