Work

NIAGARA R&D INTERNSHIP – COMPUTER VISION

JUN 2017 TO MAR 2018

Wrote a Python program using OpenCV that measures user desired dimension of a water bottle.

Canny Edge detection, Weiner filter and many image noise reduction algorithms were used to enhance the result.

Worked with a mechanical intern to convert the taken image of a water bottle into a rough 3D CAD file by modifying the FreeCAD library.

Built complete GUI for the programs using Python TkInter

Side Projects

FOREST FIRE BURNT AREA PREDICTION

OCT TO DEC 2018

Used the dataset from UCI repository along with scikit-learn and other basic python libraries.

Fit the dataset which contain sensor values, coordinates and time using Regression models such as SVM, Ridge, Lasso, ect. to fit the data then Cross validated various models to find the best fit.

Working with team member to implement a fully-connected neural network model to compare with above algorithms.

FASHION CLASSIFICATION

DEC 2018

Designed a Convolution Neural Network to classify fashion items inside the FASHION-MNIST dataset

Using Keras with Tensorflow backend, I designed a 5-layer network (3 convolutional and 2 dense layers) and archived an accuracy of 91%.

Other related work includes data visualization and preprocessing using scikit-learn.

SANFRANCISCO HACKATHON: POLICE ASSIST CAMERA – BEST HARDWARE PROJECT

MAR 2017

Programed a normal webcam and Raspberry Pi 3 to consecutively capture and scan for license plates using Python and OpenCV library.

Trained the software to recognize different type of license plates from different countries, states.

The software would print 5 possible license plate strings with percentage (ex: 2 can look like Z, 4 can look like L depend on the font and distant) in a .txt file.

Worked with fellow engineer to compare the results in the .txt file with the database that he built using mySQL. And display the license plate information such as if it’s stolen or not, owner, color, make.

The project is used to assist police quickly identify stolen car or fake plate without having to pull over or manually enter the license plate by hand.

HACK POLY – VISUALLY IMPAIRED ASSISTANCE

FEB 2018

Using OpenCV and Python, built an object detection using a pre-trained MobileNets SSD model that would detect multiple objects in the frame (RCNN) and mark their locations.

Combining code with teammates to have the program reading out loud the location of each object.

If a person is detected, perform facial recognition and tell the user who it is.

Key Skills

Machine Learning
Python
scikit-learn
tensorflow
C++
Matlab

Education

SAN JOSE STATE UNIVERSITY

SEPT 2018 – JUN 2020

Major in Master of Science: Computer Engineering

GPA : 3.6

CALIFORNIA STATE POLYTECHNIC UNIVERSITY, POMONA

SEPT 2015 – MAR 2018

Major in Bachelor of Science: Electrical Engineering

GPA : 3.75

SANTA ANA COMMUNITY COLLEGE

SEPT 2013 - JUN 2015

Graduated with AA and AS degrees in Mathematic, Physics and Liberal Arts