July 2023 - July 2024: CVML Data EngineerJuly 2024 - Present: Risk Analytics Data Scientist

I applied model monitoring techniques to current models in deployment. I leveraged tools such as Tableau and Databricks to generate automatic quality checks.

I collaborated with engineers, business stakeholders, UX designers, and clients to develop supervised classification models with the goal of grouping Deere's machine notifications via text analysis. This project assists in Deere's client webpage (OpsCenter), where clients can view machine status and notifications I used many natural language processing (NLP) techniques and machine learning algorithms to build my models. These models included SGD classification and multi-label classification. I was able to achieve upwards of 95% accuracy on predicting notification categories. The models are being put into use to allow easier functionality of notifications on OpsCenter.

Laboratory for Computational - Anthropology and Anthroinformatics(LCA)I analyzed ancient archeological sites in Peru, South America via machine learning and predictive modeling. Through this analysis, I developed methods to forecast future archeological sites in the region. This is particularly helpful now since COVID-19 restrictions have prohibited archeological excavations and explorations in these areas.We used these forecasts to develop an agent-based model with the goal of tracking the exact steps the first South Americans took in their migration routes to Peru. For example, if we set a starting point in Panama City, what would be the migration route taken to reach the coast of Peru? The model utilizes global density distributions and probability matrices to calculate raster probabilities. We are looking to expand it's starting point to other parts of the western hemisphere, all the way up to Alaska!

I explored and analyzed raw TV Media data for Abbott Ensure products. Through practices of data manipulation and formatting, I was able to develop a proof of concept that we can directly measure tv commercial impact on sales. I developed baseline models via time series modeling and quartiles with the goal of distinguishing promotional sales from tv media sales. I then extracted insights and gave examples of how we can compare and analyze different combinations of tv ads to maximize sales.

Was a software developer on a 5 week project focusing on solving problems in the sports and entertainment industries relating to the Covid-19 Pandemic.Our project was called "Hoosier Up!" and involved the creation of a virtual camp that would better connect fans and kids to their favorite athletes. We developed both a website and a prototype of a mobile app. I created our product demo which can be seen below!

Worked with Drs. Arthur Bradley and Donald Miller for eight weeks to analyze real-world cell data that was attained in their imaging lab. With the goal of better understanding colorblindness, I was tasked with finding unique characteristics in these retinal cells (i.e. short, medium, and long cells). I performed algorithmic analysis and found subtle anatomical differences between the cone types, and I ran algorithms such as neural networks and K-means clustering to construct a learning method that can help distinguish the cones for future data.My project was later picked up by a graduate student George Ji, who later published it in an ARVO journal here: https://iovs.arvojournals.org/article.aspx?articleid=2779619