MaeLSTROMS: Machine Learning for Speech-to-Text Recognition Of Master Scans

An end-to-end pipeline that ingests clinician-narrated ultrasound exam findings videos and uses speech-to-text models to produce time-coded audio transcripts, machine translation to generate foreign language-localized text, and text-to-speech models to produce localized audio.

USS-CDR: Unified Semantic Search for Content Discovery and Retrieval

The SonoSim content search engine, which uses various forms of NLP for full-text semantic search.

SonoSeg: The SonoSim Ultrasound Volume Visualization and Segmentation Platform

An internal tool used by SonoSim’s content team as part of the content creation pipeline to visualize, explore, and segment 3D and 4D volumetric ultrasound data encoded in a proprietary SonoSim data format. Based on the open-source MITK workbench, written in C++.

SR-Ultra: Deep Learning for Ultrasound Image Enhancement

A self-supervised deep residual super-resolution neural network designed to upscale and enhance 3D and 4D volumetric ultrasound data through a novel method known as salient signal recovery.

Teofilo E. Zosa, Matthew Wang, and Eric Savitsky. "Deep Learning for Ultrasound Image Enhancement." CRESST Conference 2018, Oct 1-2, 2018. [POSTER]

CI Docker Images

Pre-built containerized base environments for CorSAIR projects and CI pipelines.

GitLab CI Pipeline Templates

GitLab CI pipeline templates for use in CorSAIR-standardized projects and beyond.

cookiecutter-cruft-poetry-tox-pre-commit-ci-cd

SonoSim CorSAIR-standardized Python project creation and synchronization.

structlog-sentry-logger

A multi-purpose, pre-configured, performance-optimized structlog logger with (optional) Sentry integration via structlog-sentry.

Academia

(2017-2019)

PhD Research Mastery Exam

A Master’s thesis as well as a PhD research prospectus, with the subject area chosen specifically to best guide the efforts at SonoSim, Inc. in enhancing the SonoSim® Ultrasound Training Solution and the overall SonoSim experience.

Teofilo E. Zosa. "Catalyzing Clinical Diagnostic Pipelines Through Volumetric Medical Image Segmentation Using Deep Neural Networks: Past, Present, & Future." UCSD Computer Science and Engineering PhD Research Mastery Exam, June 7, 2019.*

* Survey paper ranked "at the level of a good paper or presentation at a top conference in the area of the exam", the rarest and highest possible examination rating.

Parallel Computation at Scale: High-Performance Computing using OpenMPI

An OpenMPI implementation of the Aliev-Panfilov electro-cardiac simulation algorithm that ran on the COMET supercomputer at the San Diego Super Computer Center.

Teofilo E. Zosa. "HPC Programming: Scaling Aliev-Panfilov Electrocardiac Simulation Using OpenMPI." CSE 260: Parallel Computation, 2018. [UNPUBLISHED]

Marvel Vs. DC: Latent Factor Models for Sentiment Analysis & Consumer Similarity Prediction

An application-specific Amazon.com web-scraper that performed sentiment analysis and generated a star-rating predictor from DC Comics and Marvel Comics graphic novel reviews.

Teofilo E. Zosa. "Marvel Vs. DC: Predicting Consumer Similarity Using Latent Factor Models." CSE 258: Recommender Systems & Web Mining, 2017. [UNPUBLISHED]

Dubble: The Robotic Prehabilitation Coach

A robotic physical therapy facilitator that used a personified adaptive interface and various physical gestures to increase user engagement and accessibility.

Bryce Woodworth, Francesco Ferrari, Teofilo E. Zosa, and Laurel D. Riek. "Preference Learning in Assistive Robotics: Observational Repeated Inverse Reinforcement Learning." Machine Learning for Healthcare Conference. PMLR, 2018.

WaNN: Convolutional Neural Networks in a Monte Carlo Tree Search

An agent that utilized convolutional neural networks and a Monte Carlo tree search to play the game of Breakthrough.

Richard J. Lorentz, and Teofilo Erin Zosa IV. "Machine Learning in the Game of Breakthrough." Advances in Computer Games: 15th International Conferences, ACG 2017, Leiden, The Netherlands, July 3–5, 2017, Revised Selected Papers. Vol. 10664. Springer, 2017.