I worked as a Research Assistant on Dr.He Zhu's research about multi-step Contrastive Preference Learning (CPL) framework. Our research aims to improve model alignment through iterative preference optimization and trajectory-based training, enhancing the performance of AI agents in complex environments.

In my role, I focus on implementing the CPL framework, which involves designing and executing experiments to evaluate the effectiveness of our approach. I also contribute to the development of a Python package for experimental design and evaluation in Metaverse interactive environments, enabling systematic performance analysis. I have been actively involved in the research process, from data collection and analysis to the implementation of machine learning algorithms.

Rutgers Open System Solution (OSS) is the subdivision of the Office of Information Technology of Rutgers University to incorporate and develop enterprise‑level software for the Rutgers community.

As a System Developer at OSS, I design and develop new software tailored to the needs of the Rutgers community. My role also involves maintaining and upgrading existing products, writing scripts for senior system administrators, and providing technical support.

One of my key projects at Rutgers Open System Solution (OSS) is When, a centralized calendar service used by 70,000+ users. When provides synchronized calendars for university-wide event sharing and enables work-study students to manage their schedules.
I led the end-to-end full-stack development, designing the front-end interface with React, building backend architecture with Golang, integrating Google Cloud APIs, implementing a MongoDB database, containerizing with Docker, and securing authentication with Rutgers' Central Authentication Service (CAS) API.
Beyond technical development, I oversaw the project from initial analysis and problem-solving through conceptualization, feature design, testing, and debugging. As a university-level product, When demanded adherence to professional standards, including collaboration with senior administrators, ensuring security and privacy protocols, and maintaining comprehensive technical documentation. The product is currently in the testing and evaluation phase.
In addition, I have contributed to maintaining and enhancing various system utilities across Rutgers, focusing on UI design, testing, and debugging. Some of these projects span over a decade, involving complex architectures, large-scale databases, and diverse languages and frameworks, requiring extensive teamwork.
I also handle product inquiries, write support scripts for senior system administrators, and participate in hiring and training new developers.

J. Kim
email: iiher@naver.com

This project is a collaborative, industry-level initiative developed in partnership with professionals from DXC Technology. The objective is to create the AI-Enhanced Customer Interaction Assistant for e-commerce platforms.

The tool is designed to optimize customer interactions by offering real-time assistance, personalized product recommendations, dynamic order tracking, multilingual support, and visual search functionalities.

This project is part of the Break Through Tech AI Studio program and is aimed at delivering a practical, market-ready product for DXC Technology. The project culminates in a formal presentation to the company, showcasing the product’s capabilities and potential applications.

• Develop an advanced AI-powered tool for e-commerce platforms using Multi-AI-Agent Framework LangGraph to enhance customer interactions by providing real-time assistance, personalized product recommendations, dynamic order tracking
• Optimized prompt engineering and hyperparameter tuning across multiple LLMs while accommodating chat history
• Developed Frontend Chatbot interface with Python-based UI framework, Mesop, allows real-time communication and updates
• Led a remote team of 4 in architecture design, agile project management, progress tracking and final integration
• Served as the primary point of contact for industrial professionals, contributing in meeting coordination, progress report & feedbacks

Check out the codes in GitHub Public repository

Watch a live demo

HarmoniHealth offers a safe, secure space where all individuals can share their stories and emotions freely. The app focuses on building relationships with individuals that have similar interests and backgrounds by pairing them up based on answers they have inputted through a form through recommandation algorithm The goal is to connect with individuals of similar experiences-- to build a support system they may not have and to remind them that they are not alone.

This project is part of HackRU event hosted by Rutgers Univierty.

• Developed HarmoniHealth, a secure mental health app using LLM to match users with companions based on shared experiences.
• Built a Spring Boot app with OpenAI API and React frontend, winning Best Design Award at a 400+ participant hackathon.
• Led a team of 4 in project concept development, role coordination, full-stack development, final integration and presentation in 24 hours.

The project is open-sourced and maintained through Github. Check out its GitHub repository

Watch a live DEMO

Macrocorr is a Python package for analyzing the correlation between macroeconomic indicators and time-series data.

• Self-developed Python package for analyzing the economic correlations between macroeconomic indicators, including 88 datasets from 265 countries
• Applied Data analysis tools including Pandas and NumPy to process and clean data, and Matplotlib for visualization
• Maintained and organized on Github with full detailed documentation and catalog

Check out the codes in GitHub Public repository

Published on Pypi and available to download Check out

As a Junior Intern at EducationUSA during the COVID-19 pandemic, I focused on adapting to new challenges and finding innovative solutions to enhance access to U.S. higher education, particularly for students in Northeast Asia and the Pacific.

My responsibilities included data analysis and assisting in policy development, utilizing technical tools and virtual platforms to conduct meetings, and leveraging social media to disseminate information..

• Design posters and social media posts for the official EducationUSA page & public events
• Assist with EducationUSA programming
• Maintaining the platform for EducationUSA officiers
• Plan, set up & publicize events, including writing speeches, creating handouts, planning activities, etc.
• Research & Collect open resources on numerous topics
• Data organization & review, specifically in Notion page
• Develop Youth outreach; Provide assistance to students who are looking for U.S education
• Build virtual meeting spaces at Gather for holding virtual events as well as providing spaces for advisors across the world
• Assist regional advisors in daily contact & event follow-ups

Vincent Flores - Lead advisor in Education USA NEAP
email: vflores@educationusa.org

As a Junior Developer at NinefoldTech, my primary responsibility is developing features for Android mobile applications, with a strong emphasis on designing core algorithms, implementing frameworks, and debugging.

I have gained valuable experience working with Mediapipe, an AI framework for tracking streaming image and video data. I applied this technology specifically for facial recognition tasks and utilized Android's camera functionality, including components like screen view and camera view, which involved 3Danalysis.

My development work is primarily conducted in Android Studio, with Java as the main programming language. I have hands-on experience in developing and deploying applications on mobile devices, which includes debugging, reading logs, and testing across multiple Android devices and versions.

In addition to mobile application development, I have experience collaborating with teams in China. I took on the responsibility of conducting research using English-based resources and providing relevant information to my Chinese team members. This also involved remote collaboration through GitLab.

J. Kim
email: iiher@naver.com

Nnew Jersey Transit is the public transportation system serving the state of New Jersey. It operates bus, light rail, and commuter rail services. It is one of the largest transit systems in the United States, providing millions of rides annually.

The EPR (Enterprise Resource Planning) system team under IT department is responsible for managing resources, including human resources, financials, supply chain, project management, and more.

Fun and casual game that is going to challenge your hand and brain! It uses a unique control feature- tilt the device to roll the ball! Colorful graphic, addictive game mode, 7 different types of questions. Enjoy the fun while becoming a math expert in no time.

• Self-studied C# language & functions in Unity platform
• Self-designed & developed this game with a comprehensive UI system, game mechanics, scenes and all other features
• Incorporate advanced Unity Package including AI, physical engine, etc
• Implements API for data management, multiplayer platform
• Published in Google Play Store with Unity ADs;

Rutgers Open System Solution (OSS) is the subdivision of the Office of Information Technology of Rutgers University to incorporate and develop enterprise‑level software for the Rutgers community.

As a System Developer at OSS, I design and develop new software tailored to the needs of the Rutgers community. My role also involves maintaining and upgrading existing products, writing scripts for senior system administrators, and providing technical support.

Technical Leadership
• Developed "When," a centralized calendar service used by 70,000+ users at Rutgers Open System Solution (OSS).
• Facilitated synchronized calendars for university-wide event sharing and work-study schedule management.
• Led the end-to-end full-stack development, designing the front-end interface with React, building backend architecture with Golang
• Integrated Google Cloud APIs, MongoDB database, Docker and Secured authentication via Rutgers' Central Authentication Service (CAS) API.
• Oversaw the project lifecycle from initial analysis and conceptualized problem solving, to agile project management and development, testing, documenting and debugging.
• Collaborated with senior administrators and cross-functional team to meet security, privacy, and usability requirements.
• Maintained comprehensive technical documenting for both internal and external usages
System Utilities Maintenance
• Maintained and enhanced long-standing system utilities, some spanning over a decade.
• Improved UI design, conducted testing, and performed debugging for complex architectures and large-scale databases.
• Worked across diverse languages and frameworks in a team-oriented environment
Additional Contributions
• Responded to product inquiries and wrote support scripts for senior system administrators.
• Played a key role in hiring and training new developers.

Jack Chen
email: jcc@oit.rutgers.edu

Overview

Research Intern at the University of Oxford, Gamification in Education Research Team, Oxford Scholastica; Our team analyzed the benefit of gamification and studied the concept of game design. We applied the aspects of gamification in developing an educational game.
“Bit's Bites” remains different from traditional quiz-style educational games, instead it incorporated basic calculations into its game mechanics. Detail can be review in the published paper.

Achievements

• Investigated the application & benefits of gamification through literature review and research from online
• Worked virtually with international scholars from all around the world
• Designed the educational game “Bit’s Bites” about mathematical operations
• Published a comprehensive Research Paper Bit's Bites: A Digital Educational Game in Oxford Journal
• Team’s representative speaker at the final ceremony

Learn more

• Check out my published paper Bit's Bites: A Digital Educational Game in Oxford Journal.
• Contact me for my Certificate of Internship & Letter of Recommendation from the professor

WALLST is an advanced personal investment portfolio review tool designed to provide users with a comprehensive analysis of their investment portfolios. Leveraging artificial intelligence, WALLST offers tailored recommendations based on user-defined parameters such as risk tolerance, investment focus, and holding period.

WALLST is committed to democratizing access to personal finance tools, empowering individuals to make informed financial decisions without unnecessary complexity or cost.

This project is built during Hackher 2025 within 24 hours. This project is awarded as Best Financial Solution of 171 participants.

• Learned, comprehended, and effectively utilized the Next.js framework within 24 hours
• Independently designed and implemented the entire frontend of the application including main color theme, UI style, flow etc.
• Successfully integrated dynamic and interactive data visualization elements to enhance user engagement and provide key financial insights.
• Utilized AI to deliver tailored portfolio recommendations, including prompt engineering, API integration, and output optimization for accurate and actionable results.
• Led a team of three, overseeing project management, coordinating tasks, and taking responsibility for the final integration and deployment of the application.
• Conceptualized and developed the project proposal and mission statement, aligning the team’s efforts with a clear vision and purpose.

The project is open-sourced and maintained through Github. Check out its GitHub repository

Watch a live DEMO