I am currently a computer vision and machine learning engineer with a cool company called Tortuga, where we develop technology to help make agriculture more efficient. Over the past nine years, I have worked with languages and frameworks including Python, Caffe, Spark, PyTorch, Tensorflow, and C/C++. Some highlights from my experience are below, but please check out my Github for a more complete list of finished projects.
Inspired by Google's Building for Billions, an initiative to make mobile appls more accessible to diverse audiences, I am completed a thesis at the University of Pennsylvania proposing a system that uses a probabalistic learning model and active learning to optimize mobile apps for both user experience and resource consumption. Alongside Neelay Velingker, Professor Mayur Naik, and Professor Osbert Bastani, we developed an effective model for understanding mobile interfaces. I have never had a more rewarding academic experience, and I owe that all to our amazing team.
Update! I recently found out that I was awarded the 2019 University of Pennsylvania Oustanding Graduate Research Award for PolyDroid! Thank you for your support!
I was fortunate to have been given a tremendous opportunity at Tortuga to manage the entire off-robot data stack, including ingestion, engineering, and machine learning. As the subject matter expert in those spaces, I designed an end-to-end microservices backend architecture that generates, validates, and evaluates deep learning models. By leveraging cloud and open source technologies, we have been able to quickly improve the robustness and learning capacity of our models without sacrificing embedded system performance. I am so happy to work with such an amazing group that has given me a great deal of latitude to explore the intersection of CV and agricultural robotics. Below are links to the classes that got me started.
In 2018-19 when I was heavily researching classical vision and deep learning theory under the incredible mentorship of Professor Jianbo Shi, my friend Peyman Norouzi and I tried to find a vision problem that had immediate tangible value. We found the Human Protein Atlas Image Classification Kaggle competition, which asked researchers to classify 27 proteins in thousands of images of cells. This problem required a lot of creativity on our end, from writing a custom network, to implementing focal loss, to devising fast original segmentation techniques. I have presented our work in multiple settings over the past year, so to learn more about this work check out my presentation slides below. Or you can view the network itself on Google Colab.
When I arrived at Accolade as a data engineer, the company was transitioning to a new enterprise data warehouse on AWS. Under the guidance of Matt Wurst, the best software manager I have ever worked for, I implemented a scalable pipeline which could collect unstructured data from microservices using SNS/SQS, transform data using Spark on EMR, and upsert the data to the warehouse in Redshift. It was a great opportunity to learn at a fun and energetic company. It meant a lot working on a team that faces the challenges of healthcare head-on and tries to make the industry a more compassionate and economical place. Learn more about them at the link below!
Accolade »Before PolyDroid, I worked with a tremendous team under the guidance of Professor Mayur Naik that focuses on debloating large programs. I researched the application of the team's Chisel program on web apps (specifically on the impact of non-determinism related to programs that interact with a server). You can read more about Chisel through the link below.
Chisel »A friend and I developed a platform to help users experiment with applying machine learning models to the cryptocurrency market. We hope to familiarize this exciting marketplace with a larger audience!
I had a lot of fun making this website, and I would be happy to help anyone get started making their personal site. Feel free to email me any questions you might have. Also, let me know if you see any ways that I can improve the site, I update it all the time.
View Code »If you have the opportunity, please take some time to check out the awesome work of some of the organizations below! Each organization supports a unique cause that means a great deal to their local communities.
My participation with Big Brothers Big Sisters programs provided me with some of the most rewarding experiences of my life. Most recently I worked as a Big Brother in Mercer County. BBBS Mercer provides opportunities for kids from diverse backgrounds in need of a role model, and they have countless success stories. Please take some time to review their recent initiatives.
Many children in urban public scools crave exposure to Science, Technology, Engineering, and Mathematic academic opportunities, but often don't know where to channel their enthusiasm. Minds Matter aims to change that for teenage students in Philadelphia through mentorship and eduction.
During our field missions in Korea, my company was often away from our base for weeks at a time. The generous people at Operation Jersey Cares sent multiple care packages to our unit during my two years overseas that provided many essential items and a little bit of home. This great organization supports deployed troops around the world and I was fortunate to see the impact of their mission first hand.
Here are some fun memories that might help you get to know me a little better!
If you would like to learn more about these destinations or activities, please reach out to me! I love trading ideas for a new adventure!
Photo editing by the very talented Marie Simonova. Check out her amazing work here!
Please feel free to contact me by phone, email, or old-fashioned mail!
But please don't show up at my apartment!
