Travis Frink Email & Phone Number

Cranston, Rhode Island, United States

Kingston, Rhode Island, United States

• Primary contributor to an ML pipeline using GuildAI, YACS, BlenderProc, and Detectron2, designed to automate and manage a large volume of computer vision-related ML experiments.• Sole developer of a deep learning-based, AI-assisted instance segmentation labeling system using OpenCV and PySOT (object tracker).• Thesis: Enhanced the "Deep Template-based Object Instance Detection" architecture (available at: https://arxiv.org/abs/1911.11822), a meta-learning-based approach to few-shot object instance detection. Improvements included advanced data augmentation for template images, architectural modifications in the backbone CNN, and a refined optimization method featuring a modified loss objective.

Providence, Rhode Island, United States

• Primary developer of a multimedia Android application for reading information, watching videos, and tracking goals related to dementia care giving.• Collaborated with clinicians and domain experts to curate and integrate specialized media and reading materials into the app.• Developed the app using the MVVM architectural pattern and AndroidX libraries. Key technologies employed include Room for database management, RxJava for reactive programming, Glide for image loading, and Gson for JSON processing.• Developed a suite of tests using Espresso, ranging from unit tests to extensive end-to-end tests, to validate various functionalities of the app.• Enhanced an existing showcase library for in-app help by integrating new styling options and enabling interactions with HTML content within WebView elements.

Kingston, Rhode Island

• Designed and developed Android app and Wear OS companion app for monitoring the walking of peripheral artery disease patients.• Developed and involved in user testing of a web application (Javascript, PHP, PostgreSQL) used for evaluating dementia patients in collaboration with Rhode Island Hospital.• Wrote Arudino software to control a SIM900 GSM module for sending automated SMS notifications.• Implemented Android app for monitoring temporal wearable sensor data for clinical settings resulting in a conference publication.

Kingston, Rhode Island, United States

• Built and tested combined Android and Wear OS applications for collecting, storing and uploading watch sensor data.• Used Docker with Apache to host user portal for accessing sensor data.• Developed user portal using Javascript, PHP, Docker and PostgreSQL for storing, managing and visualizing sensor data on the web.• Used d3 and FullCalendar to visualize sensor data.

Master Of Science - Ms, Electrical Engineering

Bachelor Of Science - Bs, Electrical And Computer Engineering