About Me
I'm currently an undergrad at Carnegie Mellon University studying
computer science, statistics and machine learning.
Experience
Machine Learning Engineering Intern
June 2024 - August 2024
Epirus
Engineered real-time drone simulations in Python and used PyTorch to train a reinforcement learning agent with Proximal Policy Optimization to optimally engage unknown drone swarms, leading to significant improvements over baseline methods.
Full Stack Engineering Intern
May 2023 - August 2023
Beaver Health
Developed a generative AI dialogue framework using GPT-4 and TypeScript to digitize evidence-based health interventions for Alzheimer's patients. Backed by National Institute on Aging & Harvard Innovation Labs.
Researcher
July 2022 - May 2023
University of Victoria
Designed and implemented Temporal Convolutional Networks (TCNs) for predicting COVID-19 outcomes, outperforming CDC benchmarks by a significant margin in predictive accuracy and timeliness.
Paper Research Poster, Slides
Creator & Software Engineer
December 2021 – May 2023
COVIDCatcher
Developed a multimodal ML model using VGG-19, XGBoost to detect COVID-19 symptoms from cough audio and health surveys, deployed with React, Flask, and AWS.
Research Poster, Slides
Software Engineer
August 2018 - May 2022
Amador Valley Robotics (AVBotz)
Implemented real-time object detection in ROS with OpenCV and C++, improving detection speed by 40%. Automated image annotation for dataset creation using YOLOv5 and DetNet in PyTorch, saving 100+ hours of manual labor.
2022 Technical Report, 2021 Technical Report, 2020 Technical Report
Projects
CLaiM
(October 2024)
Automatically file home insurance claims after natural disasters using computer vision. Won the Hyperbolic Track at Cal Hacks 2024.
(October 2024)
Automatically file home insurance claims after natural disasters using computer vision. Won the Hyperbolic Track at Cal Hacks 2024.
COVIDCatcher
(December 2021 – May 2023)
Developed a multimodal ML model using VGG-19, XGBoost to detect COVID-19 symptoms from cough audio and health surveys, deployed with React, Flask, and AWS.
(December 2021 – May 2023)
Developed a multimodal ML model using VGG-19, XGBoost to detect COVID-19 symptoms from cough audio and health surveys, deployed with React, Flask, and AWS.
The Roast
(July 2023)
Daily personalized newsletter using GPT, Claude, and LangChain to curate daily news based on user preferences. UC Berkeley AI Hackathon 2023.
(July 2023)
Daily personalized newsletter using GPT, Claude, and LangChain to curate daily news based on user preferences. UC Berkeley AI Hackathon 2023.
Shipworthy
(June 2020)
Real-time ship simulator that uses computer vision with OpenCV to steer a ship in Unity via a physical steering wheel. Won 1st overall at HackItShipIt 2020.
(June 2020)
Real-time ship simulator that uses computer vision with OpenCV to steer a ship in Unity via a physical steering wheel. Won 1st overall at HackItShipIt 2020.
SEA
(July 2020)
Deep learning pipeline using TensorFlow and Python to identify endangered marine wildlife images.
(July 2020)
Deep learning pipeline using TensorFlow and Python to identify endangered marine wildlife images.
Ad Lunam
(July 2020)
VR space exploration game using C# and Unity with procedurally generated planets and asteroid fields. Won 3rd overall at To the Moon and Hack 2020.
(July 2020)
VR space exploration game using C# and Unity with procedurally generated planets and asteroid fields. Won 3rd overall at To the Moon and Hack 2020.
Stance
(June 2020)
Developed a machine learning app using sklearn and Flask to detect toxic comments online, using LIME to provide interpretable explanations of the classifications. Won 1st overall at Data Day Grind 2020.
(June 2020)
Developed a machine learning app using sklearn and Flask to detect toxic comments online, using LIME to provide interpretable explanations of the classifications. Won 1st overall at Data Day Grind 2020.
Contact me at
bearseascape@gmail.com
ml6@andrew.cmu.edu
Website built using HTML, CSS, and JS by Michael Li