Over the course of both semesters of my senior year, I participated in the capstone courses for my Computer Science major. This capstone course centered around a project to develop for a company that would demonstrate the culmination of knowledge attained from the major and to be able to apply it in a real world setting. The project I worked on was a quantitative trading strategy app that provides an AI model to predict stock prices for novice users, with our sponsor being a Vice President from Goldman Sachs. This project was completed in a group of 4 students over the course of two semesters, and required rigorous planning and testing.
During the first semester, we focused on developing the basic AI model, as well as a rudimentary interface for testing. We focused on gathering reliable sources of data, understanding our true future user base, and creating a model based on tried-and-true AI methods using Python. From this, we established data pipelines to access a constant flow of live data, and then fine-tuned our model to be able to show relative success in stock prediction. Then, we connected the results of this model to a basic user interface to be able to demonstrate a prototype. Finally, we established that we wanted to focus on improving both the accuracy of the model and UX of the interface for the following semester.
During the second semester, we focused on establishing a much more precise model and automating all of our tedious processes, including data collection and updating. First, we heavily researched quantitative and economic research in order to determine potentially more optimal methods for stock prediction, which yielded an algorithm change that increased our accuracy. We also focused on developing automated scripts that would automatically hit the data sources (APIs and/or websites), and then extract the tables in order to collect the specific data needed for the model. Finally, we fleshed out our UI to be more aesthetic and interactive.
This project aligns with the Security competency by improving the financial security of novice investors. In the modern-day, most financial advice related to investing in the stock market is gate-kept in order to maximize profits, and thus this solution provides users the ability to create more secure financial transactions with stocks, and not randomly investing that may place high risk to their own money. This is done by having access to a model that is trained on a vast amount of data, removing the emotional component to investment, and purely looking at the numbers, creating a more financially secure environment for novice investment.
Overall, this project provided immense value towards my academic (and potentially professional) career through the exposure to developing a full-stack AI-based web application. Academically, I was able to further my Python skills by assisting in the development of a financial AI model, while also working on automating data collection, all fully applicable to the real world. This may also potentially impact my professional career, as there is potential interest in turning this project into a start-up venture, allowing me to gain experience in building a business from the ground up, and developing code for a business need.
****There is potential interest in turning this project into a start-up venture, and thus the reports will be kept confidential****
****The poster is in progress and will be uploaded at the end of the semester****