About Me
As a senior at Kent State University at Stark, pursuing a Bachelor of Arts in Computer Science, I have embraced the opportunity to combine academic rigor with impactful, real-world experiences. As a recipient of the prestigious Choose Ohio First Computer Science Scholarship, I have dedicated myself to developing both technical expertise and leadership skills, laying the foundation for a meaningful career in technology.
Outside the classroom, I have found great fulfillment in community service and global outreach. Over the past year and a half, I have volunteered to assist two Ukrainian adults with enhancing their English language proficiency. This experience has deepened my cultural awareness and reaffirmed my commitment to building connections and empowering others. Additionally, my involvement with various non-profit organizations has allowed me to contribute my skills to meaningful initiatives, further refining my adaptability, problem-solving capabilities, and collaborative approach.
Technically, I bring a robust foundation in programming languages, including C++, Python, and JavaScript, alongside practical experience in frameworks such as Flutter for cross-platform mobile application development. This versatile skill set enables me to approach challenges with creativity and precision, consistently striving to deliver innovative and efficient solutions.
I am passionate about harnessing the power of technology to address real-world challenges and drive positive change. I look forward to collaborating with professionals and organizations that share this vision, contributing my technical acumen, collaborative mindset, and dedication to excellence.
Research Projects
Analyzing the Impact of AI Tools on Student Study Habits and Academic Performance
This project explored how AI tools are influencing student learning, especially when it comes to study habits, time management, and getting useful feedback. The focus was on how these tools can support personalized learning, adjust tests based on skill level, and give professors real-time insights during class. Students responded positively overall. Many said the tools helped them study more efficiently, leading to better grades while spending less time on schoolwork. Still, there were a few concerns. Some students worried about becoming too dependent on AI, and professors sometimes found it difficult to combine these tools with traditional teaching methods. To get a full picture, we used surveys and follow-up interviews. The data gave us both numbers and personal insights. We ran descriptive stats to explore trends in AI usage and effectiveness, and used tests like T-tests and ANOVA to see how different factors influenced student experiences. We also ran regression analysis to figure out what predicted AI adoption. For the interviews, we analyzed common themes in students' responses to understand their views on the future of AI in education. In the end, the study showed that AI has a lot of potential to improve learning, but it works best when it complements traditional methods. Students and professors alike emphasized the importance of privacy, transparency, and ongoing updates to make sure these tools continue to benefit education.
View Project Download PDFExploring AI Text Generation, Retrieval-Augmented Generation, and Detection Technologies: a Comprehensive Overview
With the rise of powerful AI systems like large language models (LLMs), text generation tools have quickly become a part of everyday life. They're being used for everything from writing assistance to chatbots and content creation. But as these tools grow more advanced, so do concerns about originality, bias, misinformation, and accountability. This project takes a broad look at how AI text generators (AITGs) have evolved, what they're capable of, and the ethical questions they raise. It also dives into Retrieval-Augmented Generation (RAG), a newer method that helps AI generate more accurate and context-aware responses by pulling in real-time information. RAG offers a way to overcome some of the limitations of traditional models, which rely on fixed training data that may not reflect current or specific information. The research also covers AI detection tools designed to tell apart human-written content from AI-generated text. These tools are becoming more important as AI writing becomes harder to distinguish. Alongside the technical side, the paper looks at the ethical challenges involved — like fairness, misuse, and transparency — and how we might address them going forward. Overall, the study sheds light on how we can use these technologies more responsibly. It points to ways we can improve detection accuracy, support ethical development, and make AI more accessible without compromising trust or integrity in digital content.
View Project Download PDFNeural Network Interpretability With Layer-Wise Relevance Propagation: Novel Techniques for Neuron Selection and Visualization
Understanding how complex neural networks make decisions is becoming more important, especially in fields where transparency and accountability matter. This project tackles that challenge by focusing on a technique called Layerwise Relevance Propagation (LRP). LRP helps explain a network's decisions by tracing outputs back to the inputs that influenced them the most. But current LRP methods often have trouble accurately pinpointing which individual neurons are really making an impact. To improve on this, we introduced a new method that makes LRP more precise when analyzing specific neurons. We used the VGG16 architecture as a case study and built neural network graphs that highlight important decision-making paths. These paths are then visualized using heatmaps, and we evaluate the relevance of selected neurons using metrics like Mean Squared Error (MSE) and Symmetric Mean Absolute Percentage Error (SMAPE). We also used deconvolution techniques to reconstruct feature maps, giving us an even clearer picture of what the network is doing under the hood. Our experiments showed that this approach boosts interpretability and helps make AI systems — especially those used in computer vision — more transparent and easier to trust. This kind of work could go a long way in making AI more reliable in real-world applications, where understanding how and why decisions are made really matters.
View Project Download PDFSmart Communities and IoT Research Presentation 2024
Investigating Stress-Induced Driving Behaviors: A Psychophysiological Analysis using Wearable Sensors
View Project Download PDFSmart Communities and IoT Research Presentation 2025
Content-Specific Virtual Reality Sickness: Validation Research
View Project Download PDFArtificial Intelligence Research Presentation 2025
Implementation of AI-Generated Videos in Education
View Project Download PDFCoding Projects
Real-Time Log Analysis Dashboard
The Real-Time Log Analysis Dashboard is a web application built with React and TypeScript that visualizes server logs in real-time. It connects to a mock WebSocket server (server.js) to receive log data, which includes a timestamp, severity (INFO, WARN, ERROR), and message. The dashboard features a filter panel to sort logs by severity, a table to display log details with color-coded rows, and a bar chart (using Chart.js) to show the frequency of each severity type. Styled with CSS for a clean, responsive design, the app demonstrates real-time data handling, state management, and data visualization, making it a practical portfolio piece for front-end development roles.
View on GitHubReal-Time Stock Market Tracker
A sophisticated web application designed to provide comprehensive, real-time stock market insights through an intuitive and responsive interface. The application leverages multiple financial data APIs to deliver accurate and up-to-date stock information, demonstrating advanced front-end development techniques and robust error handling strategies.
View on GitHubNutriChat
This Nutrition Chatbot web application designed to provide detailed nutritional information for various foods. Users can input a food item and its quantity in grams, and the application will retrieve and calculate precise nutritional details from a comprehensive database containing information on hundreds of fruits, vegetables, and meats. The application features intelligent search capabilities, including a Levenshtein distance algorithm that suggests similar food items if an exact match isn't found, and dynamically calculates nutritional values based on the specified quantity. With a clean, responsive interface built using Tailwind CSS, the chatbot provides an interactive experience where users can easily explore nutrition information, view food-specific details like calories, macronutrients, and vitamin content, and receive instant, formatted nutritional insights.
View on GitHubResume
Contact
Email: benjamincward9@gmail.com
Phone: (330) 440-3605
Student Resources
Career Essentials Presentation
An overview of essential career development topics and strategies.
View Presentation Download PDFInternship Experience Presentation
Insights and learnings from my internship experience at Nationwide Mutual Insurance Company.
View Presentation Download PDF
Feel free to connect with me!