Full-Stack Software Engineer working with Typescript and C# under the W+D (Windows and Devices) OrgWorked with the WISE ML team and the telemetry data pipeline collection team

• Managed a team of nine students to create a computer vision application that allows a user to upload a picture of a crop and identifies potential major diseases on the crop, main job was to teach and mentor students about Computer Vision topics and oversee/help with coding as well as holding weekly meetings to discuss the progress of the project • Utilized the TensorFlow Python library and a Convolutional Neural Network model to train a dataset containing over 54000 images of healthy and diseased crops and achieved a testing accuracy of around 83% to identity a major disease correctly• Used React Native to create the frontend application and made a Flask server as a REST API to send model data to and receive disease classification information from to display on the frontend application • Also helped create and coordinate workshops for Computer Vision and other Deep Learning topics for students of the overall Big Data Big Impact organization with the other project leads

• Worked under the digital banking sector with the Cards Services/Management Team to help ease convenience and functionality for users of online debit/credit cards• Used the JAVA Spring Boot Framework in order to retrieve/modify JSON data about specific cards and features on each card• Created functionality to place a button corresponding to a card feature anywhere an end user desired on the UI by adding fields to the feature JSON and implementing sorting of features based on order in the backend• Limited testing of new card features to only specific employees of a financial institution by creating a new JSON key to specify user IDs for only employees who could see and test these features and implementing these user ID checks in the codebase for each feature• Implemented stricter checks on cards and the features they contained based on the financial institutions they came from, enabling these institutions to have greater control over customization of their online cards

• Worked on utilizing natural language processing methods to help improve the classification of major topics/keywords across legal documents• Created a method to more easily group legal documents by major topics by taking Word2Vec word embeddings using Google’s Gensim library, weighting these word embeddings with the TF-IDF scores of each word, and averaging each of these weighted embeddings for each word in each document to create an overall document embedding• Utilized Latent Dirichlet Allocation to extract specific topics from these grouped documents to identify the major legal issues found• Trained and produced custom word embeddings using Gensim's Word2Vec model on a large corpus of legal documents and utilized this model to identify keywords within the legal documents I had separated and clustered