Zhao Zhang
**** ******** ***** *** ***, San Diego, CA, 92122 +1-919-***-**** adcuaf@r.postjobfree.com SUMMARY
Master of Science student from UCSD with a strong academic background in robotics and machine learning. Pursuing a full-time opportunity in robotics software engineering, reinforcement learning, and deep learning. EDUCATION
University of California San Diego - San Diego, CA Master of Science 09/2018 – 06/2020 Intelligent Systems, Robotics, and Control
• Completed coursework in Robotics (Covered topics: ROS, Reinforcement Learning, SLAM, Motion Planning, Dynamic Programming), Statistical Learning and Deep Learning (Covered topics: Gaussian Classifier, Ada-Boost, Support Vector Machine, Convolutional Neural Network in Image Classification and Object Detection), Cuda GPU Programming, and Random Process
University of North Carolina at Chapel Hill- Chapel Hill, NC Bachelor of Science 08/2014 - 05/ 2018 Geophysics
• Dean’s List Fall 2014, Spring 2015; Graduated with honor
• Completed coursework in Geological Data Science (Covered topics: Fourier Transform, PCA, SVD, Convolutional Theorem) and Physics (Covered topics: Classical Mechanics, Electromagnetism) EXPERIENCE
Human Robot Interaction Project 01/2020 - 03/2020
University of California San Diego - San Diego, CA
• Created a guidance robot that can communicate with users through voice commands and lead users to their destinations
• Combined multiple nodes including Gmapping, Opencv_apps, etc. on ROS platform to achieve SLAM and other functions
• Conducted human-centric experiments with 9 participants Reinforcement Learning Project: Mutual Alignment Transfer Learning 11/2019 - 12/2019 University of California San Diego - San Diego, CA
• Realized Mutual Alignment Transfer Learning method from existing paper through python coding in a group of 2 students
• Implemented Trust Region Policy Optimization and Generative Adversarial Nets from scratch with Pytorch
• Evaluated the method on common reinforcement learning tasks in both OpenAI gym and DartEnv, and achieved results similar to those in the paper
Robot Algorithm Design Intern 06/2019 - 08/2019
Youcan Robot - Shanghai, China
• Optimized existing data fusion algorithm and composed PID control algorithm for a commercial underwater drone
• Tuned PID parameters and solved several deficiencies in underwater performance of the drone
• Suggested and engaged in developing a fish detection feature for the drone. Collected and relabeled image data from OpenImage and fine-tuned object detection CNNs including SSD MobileNet, Faster RCNN ResNet, and YOLO nets Robot SLAM and Motion Planning 02/2019 - 04/2019
University of California San Diego - San Diego, CA
• Realized robot SLAM with both Particle Filtering and Extended Kalman Filtering on dead reckoned robot trajectory
• Solved deterministic finite state and shortest path problems with dynamic programming and label correcting algorithms
• Achieved search-based motion planning and sampling-based motion planning in 3D environment, and compared two algorithms in their computation speed performance, result quality and multiusability
• Implemented both Q-Learning and SARSA on Hill Climbing problem by C++ parallel programming Statistical Learning and Deep Neural Networks 01/2019 - 03/2019 University of California San Diego - San Diego, CA
• Modified and trained several classic CNNs such as PlainNet, ResNet and LeNet on CIFAR-10 data
• Included in group project on reproducing and modifying PointNet algorithm for 3D classification and segmentation Structural Engineer Intern 05/2018 - 07/2018
China United Engineering Corporation – Hang Zhou, China
• Developed VBA program to automatically run finite element analysis on engineering structures
• Involved in commercial project to conduct floor design in CAD with group of 20 people Senior Honor Thesis 12/2017 - 04/2018
University of North Carolina at Chapel Hill – Chapel Hill, NC
• Executed preliminary research on infra-sound signals retrieved from Wanaka Scientific Balloon
• Composed R package to perform statistical analysis on signals associated to mining explosion and thunder storms
• Derived innovative model that filtered noises and located 29 target spike shape meteor signals from 4414 candidates ADDITIONAL INFORMATION
Languages Fluent English, Native speaker of Chinese Computer Tools Python, C++, MATLAB, R, TensorFlow, Pytorch, Cuda, Linux, ROS, LaTeX, Unreal, Blender Community Activity Event Director, UNC Friendship Assoc. of Chinese Students and Scholars Interests Video Game Developing, Basketball (served as Training Director in on-campus basketball club)