Aidan Mitchell

Artificial Intelligence
Engineer

About Me

I'm a recent Computer Science grad with a focus on machine learning and AI, looking for a full-time role as a Machine Learning Engineer. I've worked on projects like stock prediction and computer vision, and gained hands-on experience during my internship with the US Navy. I enjoy tackling challenging problems and am excited to contribute to innovative ML solutions.

mitchellaidan2@gmail.com

770-712-7205

Georgia, USA

Resume Github LinkedIn

Deep Predictions: Harnessing Neural Networks for Stock Price Movement Forecasting

This project investigates the integration of Long Short-Term Memory (LSTM) and Feed Forward Neural Network (FFNN) architectures to forecast short-term market movements in the technology sector of the American financial markets. Our model aims to outperform traditional linear financial forecasting methods by capturing both temporal and non-temporal market dynamics using advanced neural network techniques and rigorous data preprocessing.

Neural Networks for Volatility Estimation in Options Pricing: An Exploration

This project aims to revolutionize financial modeling by integrating artificial intelligence with traditional methods. We combine neural networks with the GARCH model to enhance the binomial options pricing framework, focusing on improving volatility estimation. Using PyTorch for neural network implementation and 'yfinance' for data acquisition, we dynamically adjust GARCH model parameters, refining inputs for more accurate options pricing. This approach adapts to the complex nature of financial markets, showcasing AI's potential to transform traditional financial methodologies.

Residual Connection LSTM Single-Image Depth Estimation

This project focuses on improving depth estimation from a single image by integrating Long Short-Term Memory (LSTM) with a double hourglass Convolutional Neural Network (CNN). By combining the temporal capabilities of LSTM with the spatial processing of the CNN, the method aims to address the challenge of lacking explicit depth information. This approach enhances depth estimation accuracy and robustness, outperforming existing methods.

Exploration on Various Performant Techniques and Solutions for the VRP

This project investigates optimization strategies for the Vehicle Routing Problem (VRP) by comparing three methodologies: brute-force, Simulated Annealing (SA), and Genetic Algorithm (GA). VRP is a complex challenge in logistics and transportation that aims to minimize costs and improve route planning efficiency. The project explores the theoretical foundations and practical applications of VRP, focusing on effective routing solutions to enhance operational efficiency.

3D Box Packing Problem - KSU Hackathon

The project aimed to develop a program that optimizes the loading of packages into a confined space while adhering to weight limits. It considered factors such as the size and weight of boxes, the size of the loading space, and box orientations. The algorithm treated the problem as a 2D bin packing problem, adjusting layer heights dynamically to maximize space utilization.