• Leading the StockScope project as Project Manager and guiding the team to explore new implementations of novel algorithms to determine minute relationships between stocks

• Building relational graphs and visual explanations for investors of all skill levels utilizing NLP and lag time prediction algorithms

• Completely revamping the frontend and working on secure user login with OAuth with JWT web token communication and full backend integration with PostgreSQL

• Building StockScope, a FinTech solution with constructive NLP integration to create an interactive investment platform, influenced by 24/7 data streams and algorithm-based analyses on stock/ETF/crypto performance and news

• Custom sentiment analysis developed in-house and fine-tuned to provide an optimal user experience

• Marries accurate medium-long term stock prediction through MACD, RSI, and Bollinger Band predictor factors with spontaneous influences using web scraping and live news analytics, and provides advice for investors

• React.js frontend, Flask backend, Alpha Vantage/Yahoo Finance API, custom Python-based sentiment analysis, plotly/D3.js interactive charting

• Part of a team of 10 developers with a focus on building an all-encompassing MVP and expanding further to adjust sentiment-analysis and adjusting the weightage of predictors through Agile-based sprints

• Competing in the C-UAS competition and am working on software for both competition (traditional two wing) and R&D (VTOL) teams

• Building software in Rust for the Purdue Aerial Robotics Team to control the on board camera and gimbal to take photos at certain times and build dynamic logic to point towards certain targets several hundreds of feet away

• Designed, fixed, and interfaced key components of the gimbal control API within the main software suite running on the drone and built new functions and control logic to streamline camera priority queue and improved communication speeds to 100 HZ for faster data parsing and actionability within avionics

• Built an internal tool with Leaflet.js and Python to pull data directly from the PostgreSQL database live and update points in a map interface to track drone movement with coordinate points and log/iterate on accuracy of the drone's movement and targetting

• Helping with development and training of a custom YOLO/DFine target detection model for the specific target descriptions as described by the C-UAS 2026 Competition guidelines with number based distinction

• Working with R and Python to solve data-based problems and applying data science fundamentals to scalable datasets, mainly from csv and csv2 files

• Conducting advanced data organization operations and plotting/visualization on meaningful topics

• Writing documentation and code breakdowns to convey data-science techniques learned and applied for each project

• Participate in Professional Development workshops and advance knowledge through career-first interactive events

• Developing a custom-trained, validated, and tested machine learning model to analyze images from social media and news to categorize and consolidate natural disaster information in real-time to provide accurate and timely reports for first responders and FEMA to act upon

• Creating data filtering and cleaning methods for images through vectorization and HSV breakdowns for use in from-scratch Decision Tree and Random Forest algorithms with distinct hyperparameter adjustments to better fit the data and adjust for imbalance

• Developing a decision tree tracking visualization algorithm to accompany the model for direct insight into the inner workings of the image "binning" and decision tree creation process

• Mainly using data from the CrisisMMD disaster information dataset, with a goal of expanding the model to include multimodal capabilities

• Working in a team with guidance from PhD candidate Soudabeh Taghian and building off of previous publications in the field