HEE JI PARK

Software Engineer

Data science

M.S in Computer Science

Resume Contact Me
About Me

Welcome to my personal website! My name is Hee Ji Park, a recent graduate from the University of Southern California with a master's degree in Computer Science. My current interests lie in the fields of Data Science (AI/ML) and Software Engineering.


I am driven by my passion for giving back to the community and seek to pursue opportunities in these fields that allow me to make a positive impact. Currently, I am actively seeking full-time positions as a Data Scientist or Software Engineer.


If you have any questions or know of any available positions that align with my interests, please do not hesitate to contact me at heejipark.usc@gmail.com or heeji.park1222@gmail.com. Thank you for taking the time to visit my website and learn more about me!

My Certifications
SQLD

SQL Developer
Related to SQL (MySQL, MariaDB, OracleDB)

OCP 11g

ORACLE CERTIFIED PROFESSIONAL
Related to Oracle Database 11g Administrator

ADSP

Advanced Data Analytics Semi-Professional
Related to Statistical Analysis, Machine Learning, R and Data Science

ENGIEER INFORMATION PROCESSING

Related to Operating System, Database, Network

Fun Fact
12 Countries Visited
80 Albumes Listened
I've seen Aurora
100000 Cups of coffee
Resume
Experience
2023.08 - Present
The Commons XR
San Diego, California, US

Software Engineer intern

2022.07 - 2022.08
LG Uplus Corp.
Seoul, South Korea

Data Scientist intern in Language Team

2022.06 - 2022.07
Seoul National University Hospital
Seoul, South Korea

Data Scientist intern in Department of Anesthesiology and Pain Medicine

2022.01 - 2022.05
USC Viterbi School of Engineering
Los Angeles, California, US

Course Grader for CSCI585 Database Systems

2020.10 - 2020.12
Seoul National University Hospital
Seoul, South Korea

Data Engineer intern in Department of Anesthesiology and Pain Medicine

2020.06 - 2020.09
NCSOFT
Seongnam, South Korea

Assistant Research Intern in Speech Recognition AI Lab

Education
2021 - 2023
University of Southern California
LA, USA

Master of Science, Computer Science

2016 - 2017
EPITA paris
Le Kremlin-Bicêtre, France

Exchange Student, majoring in Computer Science

2015 - 2018
Sejong University
Seoul, South Korea

Bachelor of Science, Computer Science and Engineering
Honor : Magna Cum Laude
Total GPA : 4.05/4.5 (94.5/100)

2013 - 2015
Baekseok University
Cheonan, South Korea

Major in Software
Total GPA : 4.41/4.5 (99/100)

University Link
My Skills
Coding
  • Python
    95%
  • SQL
    90%
  • HTML, CSS, JavaScript
    70%
  • C/C++
    50%
I'm interested in
  • Build and Manage Database
  • Data Pipeline
  • Web Programming
  • Back-end Development
  • Machine Learning
  • Data Analytics (Deep Learning)
  • Natural Language Processing
Portfolio
Job Posting Website with Knowledge Graph
Web

Job Posting Website with Knowledge Graph

Web

Development of a Job Posting Website Integrated with a Knowledge Graph

Our project focused on constructing a comprehensive knowledge graph by meticulously collecting company data and job postings from four distinct sources, along with the effective implementation of entity linking. This invaluable knowledge graph forms the backbone of a sophisticated job posting website.

Our Achievements:

Knowledge Graph Creation

We successfully compiled an extensive knowledge graph by aggregating data from multiple sources. This knowledge graph serves as a robust repository of information, allowing users to gain deep insights into companies and job postings.

Enhanced Visualization

Our job posting website goes beyond standard listings. It provides a visually rich representation of company-related data, leveraging the knowledge graph for a more engaging and informative user experience.

Intelligent Recommendations

We implemented an intelligent recommendation system that suggests similar companies and job postings to users, based on various attributes such as required skills, educational qualifications, and job keywords. This feature enhances the user’s journey by presenting relevant opportunities aligned with their interests and current viewing preferences.

Our project represents a significant step towards revolutionizing the job posting landscape, offering users a holistic and data-driven approach to job searching and company exploration.



Preview the Website!



Development Environment

  • Flask
  • Python
  • RDFlib
  • HTML5
  • CSS
  • JavaScript
  • SPARQL
  • Neo4j



Details

To access additional information and project results, please visit my GitHub repository.
Github

View Project
Applying Extractive Summarization to Long Government Documents
NLP

Applying Extractive Summarization to Long Government Documents

NLP

Applying Extractive Summarization to Lengthy Government Documents Natural Language Processing (NLP)

In this research, we employ an extractive summarization model, referred to as the GLCES model (as introduced by Xiao and Carenini in 2019), on a dataset comprising extensive government documents named GovReport. Our primary objective is to explore the challenges posed by the documents’ length and their unique domain. We perform a comparative analysis of four types of summaries:

  1. The primary summarization model, GLCES.
  2. A baseline summarization model, SumBasic.
  3. Extractive summaries generated from ground truth labels.
  4. Abstractive summaries from GovReport’s ground truth data.

To assess the quality of these summaries, we employ ROUGE scores and conduct a human evaluation, using a sample of 25 documents. Our findings reveal that the GLCES model surpasses the baseline model in terms of relevance and informativeness. However, when compared to the ground truth extractive and abstractive summaries, the GLCES summaries exhibit shortcomings in terms of consistency and coherence.



Report



Download files

To access additional information and project results, please refer to the PDF document or visit my GitHub repository.
PDF / Github

View Project
Portfolio Creation Website
Web

Portfolio Creation Website

Web

Create Your Personal Portfolio Website

Our platform enables you to craft your unique portfolio website, allowing you to showcase your skills and accomplishments.

How to make server

Technologies Used
  • Node.js: Used for managing user requests, data storage, and retrieval.
  • MySQL Database: Stores user data, including evaluations, login credentials, and input information.
  • Web Development: Our web pages are built using HTML, CSS, and enhanced with jQuery to facilitate data interaction with the server.

Preview the Website


Features

  1. Browse Others’ Portfolios: Explore other users’ portfolios on the main page.
  2. Easy Sign-Up and Sign-In: Register or log in with your SNS (e.g., Facebook, Naver) credentials.
  3. Personalized Page and Portfolio: Upon signing in, you can access your personal page and create your own portfolio.
  4. Profile Customization: Modify your profile picture and register your email in the “My Page” -> “Edit My Information” section.
  5. Portfolio Creation: Construct your portfolio in “My Page” -> “Edit My Portfolio” by:
    • Selecting a theme (choose from four options: basic, white-style, black style, simple style).
    • Adding your technical skills, career history, awards, and more.
    • Uploading your CV and showcasing your works.
    • Leaving comments on others’ work.
    • Downloading your portfolio in PDF and other formats (HTML, CSS).


Development Environment

  • Node.js(express)
  • Mysql 5.7
  • bootstrap
  • multer
  • passport

Project Setup

  1. Install necessary dependencies:
    • $ npm install
    • $ npm install -g babel-cli
  2. Modify the config/config.js file to suit your needs.
  3. Run the project initialization script:
    • $ npm run init

———————— OR ————————

  1. Install necessary dependencies:
    • $ npm install
    • $ npm install -g babel-cli
  2. Modify the config/config.js file to suit your needs.
  3. Create an “uploads” folder at the project’s root and add schema files to the “schema” folder:
    • $ mkdir uploads
    • $ schema/*.sql

Start server

  • $ npm run server

View build

  • $ npm run build



Details

To access additional information and project results, please visit my GitHub repository.
Github

View Project
Digital Trend Analyzer
Data Analytics

Digital Trend Analyzer

Data Analytics

After graduating “Magna Cum Laude,” I took the course “Big Data Analytics with Deep Learning,” supervised by the Korean government, and worked on two projects.

I undertook a comprehensive project aimed at forecasting online shopping trends using extensive data from Lotte Online Shopping. Leveraging six months’ worth of raw data, encompassing session records, purchase history, search queries, and external sources like Weather APIs, I formulated an online preference index and constructed a predictive sales model. To enhance data analysis, I employed agglomerative clustering to categorize similar products and applied regression models to anticipate sales volumes within these groups. My research also extended into network analysis, enabling me to identify patterns of product co-sales, consequently informing a robust online marketing strategy.

Challenges initially arose while dealing with the magnitude of the dataset and refining search queries. However, I overcame these hurdles by developing a novel algorithm. This experience led me to appreciate the significance of data preprocessing, as approximately 80% of my work involved natural language processing and data refinement. Ultimately, I found the process of deciphering the online membership landscape and forecasting shopping trends through data analysis to be intellectually stimulating and rewarding.

Report (only Korean)

Details

To access additional information and project results, please refer to the PDF document or visit my GitHub repository.
PDF / Github

View Project
Predicting Korean Movie Audience Sizes
Data Analytics

Predicting Korean Movie Audience Sizes

Data Analytics

Predicting Korean Movie Audience Sizes

I embarked on a project aimed at forecasting the audience sizes of Korean movies. This project was grounded in a comprehensive dataset encompassing five years’ worth of information on Korean films, complemented by data extracted from Twitter. My approach involved employing natural language processing techniques using NLTK to perform sentiment analysis on social media data. I further developed predictive models utilizing Artificial Neural Networks (ANN), Support Vector Machines (SVM), and Decision Trees.

The results of this project were indeed satisfying, as they exhibited impressive accuracy and provided meaningful insights that aligned closely with actual data. To illustrate, a Korean film titled “Hit-and-Run Squad” was initially expected to draw an estimated 5 million viewers, while our models predicted an audience of 1.5 million. In a remarkable validation of our methodology, the actual audience turnout was approximately 1.8 million viewers, a close match to our predictive model.

This project proved to be highly valuable, enabling me to rigorously assess and verify my analytical techniques against real-world data. Looking ahead, I aspire to take this project to the next level by leveraging AI analytics to delve into the realm of movie research. My vision is to analyze various aspects such as movie scenes, music, venues, and the prominence of lead actors to explore the most influential elements within each genre.



Development Environment

  1. Language : R, Python
  2. Development Environment : R studio, Anaconda Jupyter Notebook, PyCharm
  3. Library
R Python
Psych / kohonen cluster / NbClust ggplot2 / gridExtra scales / QunatPsyc caret / Hmisc PerformanceAnalytics corrplot / gplots RColorBrewer / d3heatmap GPArotation / matplotlib datetime / NumPy pandas / seaborn requests / selenium nltk / konlpy wordcloud sklearn sklearn.preprocessing sklearn.linear_model sklearn.tree sklearn.ensemble sklearn.svm sklearn.neighbor sklearn.neural_network sklearn.model_selection
  1. Data analytics :
    EDA, Crawling, Linear regression, Classification, Sentiment Analysis, SNA(Social Network Analysis), Word-cloud, Visualization


Report (only Korean)



Details

To access additional information and project results, please refer to the PDF document or visit my GitHub repository.
PDF / Github

View Project
Sejong Start-up Idea Competition
Idea

Sejong Start-up Idea Competition

Idea

SEJONG START-UP IDEA COMPETITION, Sejong University

4TH PRIZE : Pathfinding applications using AR(Augmented Reality)


Report 결과보고서 (only Korean)



Download files / 프로젝트 결과 첨부파일

Check out the files or site for more info and results on my projects.
PDF

View Project
Crazy Turtle
Game

Crazy Turtle

Game

Crazy Turtle: An Action-Packed Gaming Experience

Game Overview

In Crazy Turtle, your ultimate objective is to eliminate formidable foes, all while strategically utilizing a variety of items to withstand the relentless assault of smaller adversaries.

Game Description

  1. Inspired by the Arcade Sensation: Crazy Turtle draws inspiration from the thrilling world of arcade gaming, encompassing the dynamic interplay between large and small adversaries, in addition to the strategic deployment of items.
  2. Unique Gameplay Features: While sharing common elements with traditional arcade gaming, Crazy Turtle stands out through the following distinctive features: - Smaller adversaries actively pursue the player, adding a dynamic and challenging dimension. - Unconventional item dynamics introduce an exciting twist to the gaming experience.
  3. Diverse Item Arsenal: Crazy Turtle offers a diverse array of items that players can deploy to their advantage: - Random Teleportation Item - Invincibility Power-Up - Red Devil Boost Item (enhances speed and provides invincibility) - Enemy Amplification Item (increases the number of adversaries)

Development Environment

Crazy Turtle is developed using Python’s object-oriented programming paradigm, employing Python Turtle Graphics. The game is structured around three primary classes:

  • Protagonist & Main: This class focuses on the creation of the main character, the implementation of obstacles and barriers, and the core gameplay mechanics.

  • Adversaries: The Adversaries class handles the generation of foes and their attack strategies.

  • Items: This class is responsible for item generation and score calculation.



Download files / 프로젝트 결과 첨부파일

Check out the files or site for more info and results on my projects.
Github

View Project