- Performed Failure Analysis on ONNX, an intermediary deep-learning framework meant to improve interoperability between different DL frameworks. Paper Accepted at ISSTA 2024- Gathered 100s of different GitHub issues within multiple ONNX repositories in order to generate a taxonomy of issues to find common sources of error.- Leveraged NNSmith Automatic DNN generator to generate thousands of Neural Networks with varying architectures and layers in order to verify accuracy and robustness of the ONNX converter between PyTorch & TensorFlow 2

Mapped the most efficient pathway for swarms of autonomous drones using Python nodes by integrating A* and Dijkstra algorithms to navigate terrain detailed in binary files developed by the mapping team.Collaborated with the obstacle avoidance team to delegate command priority for drones during navigation in the Unreal Engine Simulation.

- Assembled Snowflake SQL queries for pipelining agile delivery data from company-wide Jira issues. - Created dashboards in Amazon Quicksight to incorporate tools such as outlier detection, rolling averages, and filter toggles to present company-wide data trends to senior management to standardize agile deliveries. - Developed internal enterprise tools utilizing Slack API and AWS to streamline customer data troubleshooting

- Refined and developed MATLAB application to display embedded file structures, plot time-statistics, and verify data. Sped up process of opening large file structures from 20 minutes to 1 minute. - Improved accuracy of Hyperloop track data to better reflect real-world vacuum conditions. Resulted in data with expected randomizations to test whether the pod would be prepared to handle such loads.- Increased developer productivity by pipelining MATLAB scripts and Simulink models together.

- Constructed automated Python and MySQL scripts through Amazon Web Services like Athena and EC2 in a Linux based environment displaying advertiser web traffic, earnings, and statistics onto a Grafana based dashboard. Enabled web scraper to supply machine learning algorithm with data 24/7.- Established sample and template dynamic Amazon Quicksight dashboards to market to clients. Resulted in an advertising earning dashboard competitive with other companies like Google AdSense.- Developed a self-improving Machine Learning algorithm in Python using SVM Vectors for Natural Language Processing to classify websites as adult or non-adult. Resulted in a much more sophisticated filter for the company rather than relying on a key-word detector with an 85-90% accuracy rate in 6 different spoken languages.

• Worked in burrito assembly station during rush hour and hectic times to• Handled and prepared food to be served in the line while following health code guidelines.• Cleaned work station, eating area, and line to fulfill customer satisfaction and to ensure sanitary and safework conditions.

Wayfinder Technology develops AI enabled software systems for operating employee benefits brokerages. Collaborated with another person to make informative videos explaining the concepts and purpose of Wayfinder’s data storage. Made employee training videos instructing new employees how to use the employee interface and the capabilities of the new SQL database.