● Using huggingface transformers, created a fully functional grammar error checking model that is set to be used by teams that are posting documentation for their tools● Developed python script to clean dataset given to current chatbot model and collaborated with team to create prompt questions that increased model accuracy for solutions by 12%. ● Worked with classifying datastores to prevent loss from security breaches. The project goal of detecting and fixing incorrect or missing datastore classification reduces risk from insufficient data security applied to sensitive data and drives compliance required by regulators. Efficiently saved time for tech leads by suggesting fixes rather than making them manually compare and update their datastores.● Connected company APIs and databases to find Github repositories and infrastructure that deployed specific components. ● Developed python program to parse Github repositories and find key pieces of information such as the component owner's name and the classification of the component. Program then sent emails to the component owner with the correct datastore classifications and the notice for immediate change.● Collaborated with team to find libraries that fit the criteria of the project and were secure for company use.● Application can save enormous amount of time from security breaches as data is already updated and stored properly which in turn saves millions of dollars that would otherwise be funneled into recovering and updating lost data.

Leveraging machine learning to segment organs from images and diagnose abnormal features such as tumors. Currently focused on thyroid imaging and segmentation

● Working with big data to grade surgical assessments of aspiring surgeons● Model uses paths such as tool usage, event timing, and video clarity to reach a skill score for a particular video whilst using a weight formula to justify path dependencies● Currently attempting to change efficiency in model from Peking University.

● Worked with a NeRF model that can handle articulated objects like the human body and is capable of detecting and tracking human bodies from mp4 files● Developed python script to extract frames from 3DPW Human Posing data and create bounding boxes from image frames that are used in joint detection in human actors to accurately depict their movements● Used tools such as PyQt5, CV2, and Matplotlib to accurately display joints and detection boxes on image frames extracted from videos; Utilized Linux and Python environments to properly install packages such as detectron2, Cython, and PyTorch

● Worked with Pramila Srinivasan to design an automated chatbot to interact with customers. ● Passes in file of commonly asked questions from customers in order to provide correct answer depending on generality of customers question● Collaborated with software team and utilized Google’s Dialogflow API to connect hundreds of common customer questions to hand-written answers; Implemented API into website through Kommunicate ● Increased customer satisfaction by providing potential clients with answers to questions when moderators are unavailable

● Worked with website development team in order to create a more user friendly experience to potential customers by adding widgets and shopping cart to website● Utilized HTML and Javascript to create a user friendly shopping experience as well as advertisements on homepage of website● Developed a customer feedback service using HTML that would allow customer's to place reviews on products they bought

● Utilized Computer Aided Design Software for 3D printing of porous material for use as bone in prosthetics ● Efficiently created new patterns of pores and visualized bone structures for thousands of potential samples by placing pores at precise coordinates in bone layers using AutoCAD software● Improved bone strength those in need through careful designing and testing processe¬s of specific pores in bones through 3D printing software of structures built in AutoCAD