Marcus Lee Email & Phone Number

San Jose, Ca, Us

Ithaca, New York, United States

• Developing an AI-enabled traffic management system for advanced air mobility, creating multi-agent coordination models for adaptive drone cooperation in shared airspace; sponsored by a NASA USRC grant• Built and evaluated foundational baselines with decision transformer and multi-agent transformer architectures, establishing robust training pipelines for multi-agent environments to benchmark future advanced architectures• Optimized large foundational model training on NVIDIA DGX-H100 GPUs for multi-agent cooperative robotics during the NASA and NVIDIA OpenHackathon; leveraged CrossFormer transformer-based architecture to develop specialized navigation models, providing insights into efficient GPU-based multi-agent training

Ithaca, New York, United States

• Developed assistive robotics implementations in a full-stack robotics lab dedicated to improving the quality of daily activities for mobility-impaired individuals through collaborating on open source research projects• Adapted and implemented the BodyPressure model for processing multi-modal data to detect and infer real-time human pose estimation despite occlusions for improving informed human-robot interaction• Designed physical and Unity simulated robotics toy environments integrated with static and dynamic collisions for teleoperated data collection and for training heuristic policies of robot arm tactile skin sensor configurations

Santa Clara, California, United States

• Developed Autonomous Lab Assistant (C1CO) robot with added features including facial recognition, object detection, and two robotic arms purposed to perform high force operations and small precise movements • Trained and evaluated custom real-time object detection models using Tensorflow and Tensorflow Lite APIs using self-collected and annotated dataset of various lab equipment tools while balancing speed and accuracy • Designed a real-time kmeans image segmentation algorithm from RGB and depth frames, paired with object detection to distinguish between target objects and obstacles, facilitating seamless input for inverse kinematics

Ithaca, New York, United States

SoNIC Assistive Robotics Workshop

Bachelor'S Degree, Computer Science

High School Diploma, High School/Secondary Diplomas And Certificates

Education record