Samsung Research Global AI CenterAI Methods Team

Developed and deployed real-time objection detection system for perimeter surveillance in army ammunition depotsDeveloped a data augmentation algorithm using weather translation GAN to enhance object detection performanceProposed a novel personalized federated learning method to improve global performance of object detectors across multiple military bases

Assisted professor Frank Hutter in grading exercises of the graduate level deep learning course

Maintained the codebase of Auto-Sklearn, a python library for automated machine learningImplemented unit tests and configured a CI pipeline using Travis CI to improve codebase stabilityIdentified and fixed bugs in various components of Auto-Sklearn, significantly reducing the user-reported defects

Assisted Reto Schoelly in the creation and administration of the robot design courseInstructed students in programming with Python to control motors and light sensors using Raspberry Pi GPIO interface