At El Aula Azul, the children learn English through a method of “immersive learning” where they are provided alternative and more engaging forms of instruction like cooking, dancing, making crafts, singing, performing, and much more. While volunteering at El Aula Azul, I assisted the teacher in planning and carrying out these activities.

For the 2019 Microsoft Global Hackathon, I worked with Kids in Need of Defence (KIND) which is a non-profit that provides legal services and representation to children navigating the US immigration system. My team's main focus was their Child Portal which serves as a quick reference for the children to review their rights or the documents they need. Over the course of three days, our team optimized the Child Portal for mobile, deployed it on Github Pages, and connected it to a custom domain.

During the winter break of my sophomore year at UMass Amherst, I traveled to USVI St. Thomas to assist in hurricane relief and recovery from the damage done by hurricanes Maria and Irma. Daily, we were assigned a team and a work site. Work ranged from debris removal to sorting out belongings with the family to keeping the owner's puppy occupied. The atmosphere of All Hands and Hearts was wonderful and the people of the island were incredibly strong and amiable.

After earning the rank of Eagle Scout, I decided to continue to assist the troop as an adult leader. I still attend meetings when my schedule permits and I participate in campouts to help guide younger scouts.

Member of Cohort 3 of the Microsoft AI Development Acceleration Program (MAIDAP).

Over the course of 2 years, I will work with 4 different teams from around Microsoft with the goal of kickstarting or enhancing one of their current AI-related efforts. Each rotation, a project team usually consists of 1 Program Manager, 2 Engineers, and 2 Data Scientists. As a Machine Learning Engineer, my role is to act as a technical liaison between the analysis and modeling work of our team's data scientist(s), and the software architecture required to implement their solutions at scale.

Some tools used: Python, Pandas, Azure ML, SQL, Azure Data Explorer, Spark, Databricks, ...

More information about our work can be found here: https://blogs.microsoft.com/newengland/2018/09/26/microsoft-ai-development-acceleration/

I worked with Prof. Erik Learned-Miller and Graduate Research Student Archan Ray on a project with the goal of detecting and recognizing text in cartographic images. I performed literature review and implemented TextBoxes++, a current state-of-the-art solution, to compare with our team's results. My work can be found here: https://github.com/seangtkelley/icken-and-chegg

I performed software development work with Dr. Giuseppe Masetti, a Research Assistant Professor at the UNH Center for Coastal and Ocean Mapping. I have contributed to the front-end of his Hydroffice Smartmap GIS web application. I made use of the OpenLayers Javascript library which is written in ES6, transpiled to ES5, and minified and bundled by Webpack. I recently completed a feature that allows the user to animate the available data layers and I am currently working on a feature that would mimic the hydrographic survey planning capabilities of HYPACK or ESRI ArcMap. During this academic year, I will continue to work on this project remotely, using the GitHub comment and issue system as well as Google Hangouts to maintain open and active channels of communication.

I worked on an Airborne LIDAR Bathymetry (ALB) research project lead by an M.S. candidate, Mr. Matt Birkebak and his advisors Dr. Firat Eren and Dr. Shachak Pe’eri. My work primarily focused on assisting in Mr. Birkebak’s research of characterizing and modeling the effects of the water surface on the laser beam. At their current state, nautical charts including ALB data are not accurate enough to use to navigate. It is imperative to improve the accuracy of the data as ALB is much cheaper and more efficient than SONAR for shallow and shoreline waters. When I joined the project, one experiment had been done but there were considerable uncertainties with confounding variables. I worked with Birkebak, Eren and Pe’eri to create and conduct a more accurate experiment. The data was collected in the 120-foot Wave Tank in the UNH Chase Ocean Engineering Research Laboratory. I then preformed a considerable amount of data analysis in Matlab. Our main goal was to confirm or discount models created from the previous experiments. All code was maintained in a Git repository synced to BitBucket so it could be shared with other researchers on the project. On several occasions, I shared the correlations and patterns I found with my co-workers in the form of oral and poster presentations as well as summary papers.