Software Engineer

General Software Development United States

Profile
Contacts

Summary

Software Engineer Intern, Kinetic Vision: June 2020- August 2020
Developed a full semantic segmentation pipeline for a dataset of over 100,000 images with 54 classes.
Designed and implemented a computer vision pipeline for one of the largest retailers in the world.
Integrated an IoT sensor system, using over 1,000 hours of footage.
Collaborated with a diverse team to improve retail stocking process speed by 8%.
Android Development Tech Lead, Codepath: December 2019- August 2020
Led, coordinated, and organized 15 Android developers, creating a variety of applications.
Developed and designed 10+ Android applications, as well as helped students design over 30 applications.
Increased the largest pipeline of high-performing underrepresented engineers in tech via Codepath.
Taught Facebook-designed, for credit, 12-week Android development course on campus with 25+ students.
Software Engineer Intern, System Corp: June 2019- August 2019
Wrote an image segmentation and classification algorithm, using Python and C++ to diagnose 50 skin diseases.
Launched a full diagnosis pipeline from image reception to inference on 22 hospital servers.
Reduced diagnosis time from 2 hours to 2 seconds per patient and improved diagnostic accuracy by 4%.
Collaborated with and learned from a dedicated team of 12 data scientists and software developers, to solve diverse problems.
Performed 356 hours of applied research in machine learning, contributing to the medical research community.
Software Engineer Intern, Kashmir World Foundation: April 2019- June 2019
Developed an artificial intelligence system, in C++ and Python, to classify and detect animals in camera trap image feed, to help save snow leopards and other 12 endangered species.
Wrote a research paper on 412 hours of applied, deep learning, research conducted.
Cut down animal image classification process conducted by preservation workers from 2 weeks to 5 minutes, by implementing a full image processing pipeline.
Overcame challenges of occlusion, camouflage, and localization for a data set containing +100,000 images.