Electrical and Computer Engineer

Jianze Mao

**** ************ **** *** ***, San Mateo, CA, 94404 ************@*****.*** TEL: 619-***-**** EDUCATION

MASTER OF SCIENCE UNIVERSITY OF CALIFORNIA, SAN DIEGO

• Major: Electrical and Computer Engineering (GPA: 3.70/4.0) Sep. 2017 - Mar. 2019 BACHELOR OF SCIENCE UNIVERSITY OF CALIFORNIA, SAN DIEGO

• Major: Electrical and Computer Engineering (Major GPA: 3.826/4.0) Sep. 2013 - Jun. 2017

• Henry G. Booker Award 2016 – 2017 (for UCSD ECE students graduating with GPA higher than 3.7) SKILLS

• Programming Skills: Python, Java, SQL, C++, C, Matlab, Verilog

• Big Data and Deep Learning Framework: Apache Spark, Tensorflow, Hadoop, MySQL, PostgreSQL

• Machine Learning Models: Decision Tree, Random Forest, XGBoosting, SGD, Linear/Logistic Regression

• Deep Learning Models: Convolutional Neural Networks(CNN), Recurrent Neural Networks(RNN), LSTM

• Tools: Git, GitHub, Anaconda, PyCharm, VMware, PyTorch, AWS, Visual Studio, Vim, Jupyter Notebook

• Platform: Linux, Windows, Unix, iOS

PROGRAMMING PROJECTS

MORSE CODE ENCODER/DECODER IN PYTHON

• Established a Class with Morse Code encoder and decoder APIs based on Python fundamental data structures

• Saved the Morse Code table as key-value pairs by utilizing Python dictionary in constructor function

• Achieved a user-friendly Morse Code Translator with efficient time complexity and space complexity SAN FRANCISCO CRIME ANALYSIS ON APACHE SPARK

• Analyzed the crime data from 2003 to 2018 in San Francisco by exploring the data in terms of spatial distribution, time series, and categories of crime

• Utilized Spark Dataframe and Spark SQL to perform OLAP analysis on real crime data and visualized the results in Spark

• Drew several insightful conclusions based on the visualization and gave suggestions to SFPD CNN BASED CAR CLASSIFICATION ON TENSORFLOW

• Initialized the CNN model with “VGG19” as base model and appended with a global average pooling layer, a fully connected layer, a dropout layer and an output layer with “softmax” activation function

• Achieved 0.65 accuracy by training the CNN model when the pretrained “VGG19” was fixed and optimizing the customized layers

• Improved the accuracy to 0.85 by setting a few layers in “VGG19” as trainable and fine-tuning the whole model

• Evaluated the model by doing classification on testing data that included about 10,000 images for 196 car classes in total, and the model gave an accuracy at about 0.8

EXPERIENCES

RESEARCH ASSISTANT, UCSD DEC. 2018 - APR. 2019

• Tested the performance improvement of MFSK demodulator with 4th order Butterworth Band-pass Filter in AWGN Channel compared to the one with 2nd order Band-pass Filter

• Found the best pair of bandwidth and tone space and plotted the minimum symbol error rate corresponding to input energy by implementing the filters and running simulations in Matlab

• Achieved 0.5 dB improvement by comparing the input energy at the same level of symbol error rate

Contact this candidate