Hongjun Wu Updated on: September 22, 2023.
Github: @wu-hongjun LinkedIn: @wu-hongjun
Google Scholar: Hongjun Wu Art Station: @wuhongjun
Resume
Feel free to download my most recent resume here: Hongjun Wu’s Resume
About
I combine the power of technology with the beauty of art to create products that are both technically impressive and aesthetically pleasing.
My current research focuses on machine learning and its application in game development. I am also a trained artist working in traditional mediums of printmaking, photography, and illustration.
Education
Cornell University 2021 - 2023
Master of Science in Information Systems.
Technion 2021 - 2023
Master of Science in Applied Information Science.
University of Washington 2017 - 2021
Bachelor of Arts in Interdisciplinary Visual Art.
Valley Christian High School 2013 - 2017
Diploma with Honor, Varsity Athlete in Robotics.
Professional Experience
Brownie AI 2023 - Present
Co-Founder, Chief Product Officer
Trunk Tools 2023 - Present
Machine Learning Software Engineer
Lerer Hippeau 2023
Artificial Intelligence Engineer
Dreamhaven - Moonshot Games 2022
Technical Artist Intern
TrackAir 2022
Co-Founder
Carter Burden Network 2021
Data Scientist
Open Source Projects
Heiya GUI 2023
Independent Developer
Heiya 2022 - 2023
Independent Developer
Xwift 2023
Independent Developer
Honors & Awards
Cornell Tech Startup Award
Finalist 2022
Zumtobel Group Awards
Special Prize for Innovation (EUR €20,000) 2021
Jacobs Technion-Cornell Institute
Connective Media Fellowship (USD $30,000) 2021
University of Washington
Dean's List 2018 - 2020
FIRST Robotics Competition - Central Valley Regional
Team Spirit Award 2017
Productions
Applied Machine Learning Open Book @ Cornell Tech 2022 - 2023
Producer
SuperFly VR @ University of Washington Reality Lab 2020 - 2021
Game Artist
"Dazzle!" @ University of Washington Animation Research Labs 2019 - 2021
Lead Animator, Lead Modeler, Technical Artist
Publications
[3]. Felix Heisel¹, Vesela Petrova¹, Hongjun Wu¹, RhinoCircular: Integrated Material and Construction Circularity Evaluation Tool for Early-design Phases, SBE22 Berlin Sustainable Built Environment D-A-CH Conference, 2022.
Abstract
According to the European Commission, “up to 80% of [a] product’s environmental impacts are determined at the design phase.“ (European Commission, 2015) Yet, current building or construction circularity evaluation tools are calibrated for employment at a much later stage in the planning process, often only utilized to analyse design decisions retroactively.RhinoCircular (RC) is an early-design circularity evaluation tool for the popular CAD software Rhinoceros3D and its Grasshopper environment. (Heisel and Nelson, 2020) Specifically designed to provide immediate feedback on design decisions pertaining to material selection and construction systems within the uncertainties of the early-design phase, the tool also allows for the creation of material passports (MP) with increasing accuracy in later planning stages. To accommodate varying input requirements, RC relies on two sets of material databases. A generic database provided by RC containing common materials and material families is well suited for early-design evaluations. Through industry collaborations, RC is additionally compatible with product/producer-based databases and their set of materials and products. Based on these complementary inputs, RC then dynamically visualizes the adaptable computation of pertinent circularity values within the design environment, and allows their export into building passport platforms (BP) or in the form of industry standard MPs. RC’s material attribution and calculation outputs are further compatible with other common CAD / BIM software platforms for data and model exchange.
This paper specifically addresses the challenges of early-design evaluations through the development and implementation of a set of (adaptable) construction archetypes applicable to modelling inputs with a lower degree of specificity. This library of visual construction details connected to both data input sets allows for a design- and user-driven, intuitive and scalable feedback process in times of low-resolution modelling inputs. Through case study examples, the paper aims to evaluate the feasibility and validity of this approach with respect to circularity evaluation, data input and MP export.
[1] European Commission. 2015. “Closing the Loop - An EU Action Plan for the Circular Economy.” Communication from the Commission to the European Parliament, the Council, The European Economic and Social Committee and the Committee of the Regions 614 final. COM(2015). Brussels, Belgium: European Commission.
[2] Heisel, Felix, and Cameron Nelson. 2020. “RhinoCircular: Development and Testing of a Circularity Indicator Tool for Application in Early Design Phases and Architectural Education.” In AIA/ACSA Intersections Research Conference: CARBON. Pennsylvania.
[2]. Rina Desai, Lisa Fernandez, Dozene Guishard, Clewert Sylvester, Hongjun Wu, Tech Pals: Results from a Smart Screen Technology Pilot for Homebound Older Adults, Carter Burden Network, 2022.
Abstract
With a growing social reliance on technology to access resources, meet basic needs, and stay connected, it is becoming increasingly vital to help older adults cross the “digital divide”. In particular, homebound older adults experience greater challenges with meeting needs and accessing technology programming. In order to address this disparity, the Carter Burden Network (CBN), an aging services nonprofit organization in New York City, in partnership with the Roosevelt Island Disabled Association (RIDA), a volunteer-led community organization, launched the grant-funded “Tech Pals” technology pilot on Roosevelt Island to connect 40 homebound older adults with smart screen devices, one-on-one set up and training, ongoing troubleshooting support, and virtual programming opportunities. The smart screen device selected for the project was the Amazon Echo Show, based on features that could support the needs of older adults including voice controlled digital assistance, video calling, automatic updates, and a range of accessibility features. The goals of this project were to: improve connectedness, support independence, enhance self-efficacy, and improve quality of life for participants.In this study, we evaluated the impact of introducing the Amazon Echo Show and technology education into participants’ lives over a span of two years, using qualitative and quantitative data analysis methods. The study initially consisted of 40 participants who received a set of pre surveys including the validated General Self-Efficacy Scale, the validated Loneliness Scale, and a customized pre survey collecting data on demographics, comfort with technology, health behaviors, utilization of the Echo Show, and perceptions of the project. Participants were provided with corresponding post surveys after at least three months of project participation. In total, 33 post surveys were collected as a number of participants passed away during the course of the project or became unreachable. A focus group was also conducted with nine project participants and four caregivers in June 2021.
[1]. Taso G. Lagos¹, Yash Singh², Andrew Pace², Erik Levi Stone², Hongjun Wu², Hongyi Yan², Shayla Forbes-Luong², Studying Abroad Meets Marginalization: Roma of Greece, Autoethnography, and Academic Tourism, Journal of Tourism and Management Research, Issue 2022, Vol. 9, No 2: 125 – 139. https://doi.org/10.18488/31.v9i2.3138
Abstract
While Greece has historically hosted many minority groups in various relational statuses with the majority population, the Roma uniquely embody practical, psychological and metaphorical spaces that sets them apart from other excluded groups. This study explores the historico-social space that separates the Roma and contextualizes recent developments, including Covid-19, which further marginalizes the group. The transactional space that defines relations between Roma and non-Roma encompasses a ‘gaze’ that disenfranchises Gypsy cultural standing and reduces mutual understanding between mainstream and marginalized communities. This same transactional space is rife with misunderstanding that profits normative day-to-day relations between Roma and those in mainstream society. The paper explores perceptions of the Roma within the Greek social hierarchy, while suggesting study abroad programming, as part of academic tourism, can play a positive role in altering perceptions of minority groups.
Manuscripts
[3]. Xuanyu Fang¹, Yunzhu Pan¹, Hongjun Wu¹, Feng-Shui Compass: A Modern Exploration of Traditional Chinese Environmental Analysis, Ubiquitous Computing, Cornell Tech, 2022. https://arxiv.org/abs/2210.13672
Abstract
The technological advancement in data analysis and sensor technology has contributed to a growth in knowledge of the surrounding environments. Feng Shui, the Chinese philosophy of evaluating a certain environment and how it influences human well-being, can only be determined by self-claimed specialists for the past thousands of years.We developed a device as well as a procedure to evaluate the ambient environment of a room to perform a study that attempts to use sensor data to predict a person’s well-being score in that environment, therefore evaluating the primary aspect of Feng Shui. Our study revealed preliminary results showing great potential for further research with larger experiments.
[2]. Kehan Wang¹, Jiaxi Yang¹, Hongjun Wu¹, A Survey of Toxic Comment Classification Methods, Applied Machine Learning, Cornell Tech, 2021. https://arxiv.org/abs/2112.06412
Abstract
While in real life everyone behaves themselves at least to some extent, it is much more difficult to expect people to behave themselves on the internet, because there are few checks or consequences for posting something toxic to others. Yet, for people on the other side, toxic texts often lead to serious psychological consequences. Detecting such toxic texts is challenging. In this paper, we attempt to build a toxicity detector using machine learning methods including CNN, Naive Bayes model, as well as LSTM. While there has been numerous groundwork laid by others, we aim to build models that provide higher accuracy than the predecessors. We produced very high accuracy models using LSTM and CNN, and compared them to the go-to solutions in language processing, the Naive Bayes model. A word embedding approach is also applied to empower the accuracy of our models.
[1]. Chenran Ning¹, Hongjun Wu¹, Yue Liu¹, Deliberation in health-related headlines, Psychological and Social Aspects of Technology, Cornell Tech, 2021. https://psyarxiv.com/e5bn7/
Abstract
Just how big of a difference will deliberated thinking make in the digital age when judging whether the headline of a digital article is true or fake? Misinformation plagued the Chinese internet space, and fake news, especially related to health tips, often went viral on the internet with rapid speed. A previous study 1 was previously conducted on political articles measuring the influence of partisanship on thinking deliberately. In this paper, we conducted a study on how deliberation influenced the accuracy of Chinese netizens distinguishing real and fake news headlines, using a similar experiment procedure from the above mentioned study. We found that deliberation reduces the possibility of these readers being misguided by fake health-related headlines. A similar trend of accuracy was observed when participants thought deliberately compared to the original study, despite using different topics on a different population of participants.
Research Projects
United Nations (UNDEF) - CSO Incubation Platform 2022
Developer
Amada - Mixed Reality Heavy Machinery Prototyping 2022
XR Developer
Cornell University Circular Construction Lab - RhinoCircular 2021 - 2022
Software Developer
University of Washington Lagos Group - A Study on the Roma Refugees in Greece 2019 - 2021
Ethnographer
University of Washington Aeronautical Laboratory - Wind Tunnel Calibration 2017 - 2020
Software Developer
CERN - Deep CNN Long-live Particle & Dark Matter Detection 2019
Data Scientist
Social Involvements
Roosevelt Island Seniors Association - Web Literacy Course 2021
Instructor
University of Washington School of Art - Printmaking Association 2018 - 2021
Graphic Artist
University of Washington School of Art - LINK Space Studio 2019 - 2020
Monitoring Specialist & Technician
FIRST Robotics Competition - Team 4415 - EPIC RobotZ 2014 - 2017
Co-Founder, CEO, CNC Lead
Academic Positions
Cornell University - INFO 5305: User Experience & User Research
Grading Assistant
Professor: Shiri Azenkot Spring 2023
Cornell University - CS 5787: Applied Machine Learning
Grading Assistant
Professor: Volodymyr Kuleshov Autumn 2022
Cornell University - CS 6785: Advanced Topics in ML - Deep Probabilistic and Generative Models
Course Assistant
Professor: Volodymyr Kuleshov Spring 2022
Cornell School of Art, Architecture, and Planning - Circular Construction Lab
Research Assistant
Supervisor: Felix Heisel Autumn 2021
University of Washington - CSE/STAT 416: Introduction to Machine Learning
Teaching Assistant
Professor: Vinitra Swamy Summer 2020
Professor: Valentina Staneva Spring 2020
Professor: Sewoong Oh Spring 2019