Engineering Assistant

Qinghao Meng

Email: (school) ********@***.*** or (self) *********@*****.*** Cell phone: 206-***-**** EDUCATION

UNIVERSITY OF SOUTHERN CALIFORNIA, Viterbi School of Engineering Los Angeles, CA Master of Science in Computer Science, Expects to graduate in May 2022 UNIVERSITY OF SOUTHERN CALIFORNIA, Viterbi School of Engineering Los Angeles, CA Master of Science in Electrical Engineering, Graduated in May 2020 UNIVERSITY OF WASHINGTON, College of Engineering Seattle, WA Bachelor of Science in Electrical Engineering, Graduated in June 2018 TECHNICAL SKILLS

Programming languages: Java, Python, MATLAB, C/C++, Bash script, HTML, CSS, Javascript, SQL Tools and Framework: VSCode, Android Studio, Eclipse, Node.JS, React, Pytorch, TensorFlow, AWS, Git Operating Systems: Linux, Windows Database: MangoDB, MySQL Relevant Course Work: Programming Systems Design, Data Structures and Algorithms, Web Development, Android Development, Machine Learning, Deep Learning, Computer Vision, Probability and Statistics, Linear Programming INDUSTRIAL AND RESEARCH EXPERIENCES

The Collaboratory for Algorithmic Techniques and Artificial Intelligence Research Assistant July 2020 – now Engaged in fastmap algorithm research. The fastmap is a graph preprocessing method which embeds the RGB images to higher dimension features. We would like to know the benefits of fastmap features. Implemented the fastmap algorithm in python and applied fastmap features to image classification problem. Achieved 70% accuracy on 10 objects image classification, which is similar to the accuracy of traditional CNN. Intellectual Venture Corporation Undergraduate Capstone Co-op Project January 2018 – June 2018 Analyzed bacterial infection of human brain by monitoring the peaks in each brain signals. Developed a peak detection algorithm using MATLAB to each brain signal using sliding window algorithm and found the position and the number of peaks in each signal. Designed the GUI for user to tune the threshold of peaks. RELEVANT COURSES PROJECTS

CNN based face recognition under complex background Python Tensorflow keras AWS Final deep learning project design, Spring 2020

Collected CelebFaces Dataset which contains 202,599 images from 10,177 identities. Preprocessed face images by using facenet detectors in OpenCV library to cropped the face from each image. Applied Amazon AWS GPU server to train our CNN model and achieved 60% accuracy for celebrities identification. Kobe Bryant Shot Attempts Prediction MATLAB

Machine Learning with Signals, Fall 2019

Collected dataset from Kaggle competition and designed a supervised binary classification system to predict whether each 30,698 shots from Kobe Bryant’s career was attempted in the playoffs seasons or in the regular seasons. Preprocessed the dataset for missing data, class imbalance, data standardization and PCA for dimension reduction. Implemented logistic regression, random forest and k-nearest neighbors and fine-tuned the hyper-parameters of each classifier using 5 fold cross validation.

Achieved 0.6 F1 score and 0.8 overall test accuracy on the test dataset for our final system. Search Engine JAVA

Data Structure and Algorithm, Winter 2018

Developed a mini search engine similar to Google Search and Microsoft Bing. The users type a query and our engine can search the relevant webpages among the webpages dataset provided by instructors. Calculated TF-IDF vector of the query and websites and computed the cosine similarity between the given query and each websites, then found the most relevant webpages to the given query. Developed page-rank algorithm for the webpages dataset to assess the popularity of the websites. Combined results of cosine similarity and page-rank algorithm and selected the most relevant webpages for the query.

Contact this candidate