Education
-
Carnegie Mellon University,
School of Computer Science, Pittsburgh, PA
M.S. in Artificial Intelligence and Innovation | GPA: 3.96/4.00 | Sept 2021 - May
2023 (Expected)
Core Courses: Deep Learning (A+), Multimodal Machine Learning (A), AI Engineering (A)
Coding & Algorithms Bootcamp (A), Machine Learning (A+), Web Application Development
-
University of Michigan,
Ann Arbor, MI
B.S.E. in Computer Science (Summa Cum Laude) | GPA: 3.83/4.00 | Sept 2019 - May
2021
Core Courses: Machine Learning (A+), Advanced Operating Systems (A-), Multidisciplinary
Design Program (A), Computer Vision (A+), Linear Algebra (A+), Computer Organization (A-)
-
University of Michigan - Shanghai Jiao Tong University Joint Institute,
Shanghai,
China
B.S.E. in Electrical and Computer Engineering (Outstanding Graduates) | GPA: 3.55/4.00 | Sept 2017 - Aug
2021
Core Courses: Programming & Elem Data Structures (A+), Electronic Circuits (A),
Hornors Calculus (A-),
Intro to Computers & Programming (A-), Probabilistic Methods in Engineering (A-)
-
Technical University of Berlin,
Berlin, German
Winter Program: German (A) | Jan 2018 - Feb. 2018
-
Shanghai Kongjiang High School,
Shanghai, China
High School | Sept 2014 - Jun 2017
Sufficiently Secure Controller Area Network
Mert D. Pesé, Jay W. Schauer, Junhui Li, Kang G. Shin.
2021 Annual Computer Security Applications Conference, ACSAC 2021 [Webpage]
The Role of Monetary Incentives In Human Privacy Preference: Insights from Global RCT Experiments on
Mobile App Users
Ranjan Pal, Yixuan Wang, Junhui Li, Mingyan Liu.
The Institute for Operations Research and the Management Sciences, INFORMS 2020 [Pre] [Slides]
Privacy Risk is a Function of Information Type: Learnings for the Surveillance Capitalism Age
Ranjan Pal*, Junhui Li*, Jon Crowcroft, Yong Li, Mingyan Liu, Nishanth Sastry
IEEE Transactions on Network and Service Management, TNSM 2021 [Webpage]
Colorizing Black and White Picture
- Computer vision course project mentored by
Andrew Owens
- Reimplement and optimize the colorizer devised by Richard Zhang [2] through adjusting the architecture
of the neural network as well as designing a more effective loss function.
[Proposal] [Final Report]
Sufficiently Secure Controller Area Network
- Instructed by Mert Dieter Pesé and
Kang G. Shin
- Developed a sufficiently secure alternative CAN with minimal
overhead on resources and latency by leveraging protocol-specific properties of CAN.
[Webpage]
DoS Counter Attack by MAC Layer Bypass
- Instructed by Mert Dieter Pesé and
Kang G. Shin
- Construct novel anti-spoofing defense system with a GPIO controller against compromised
electronic control units (ECUs), which manages to avoid flooding the bus,
and has lower network latency. (image adapted from [1])
[confidentiality required]
Holistic Privacy Framework for Connected Vehicles
- Instructed by Mert Dieter Pesé and
Kang G. Shin, sponsored by
Ford Motor Company.
- Design a privacy-preserving automotive data collection and sharing architecture that effectively
anonymizes vehicle data to prevent third-party application from identifying drivers or their behaviors.
[confidentiality required]
Meta, Menlo Park, California
Applied Research Scientist Intern, Enterprise Products Dept. | May 2022 - Aug 2022
Project: Spend Classification
-
Improved and tuned a cross-lingual language model (XLM) that saved $480M+ in the last two
years and serves 99 internal teams by classifying Meta’s daily spending into 400+ invoice
categories with ~$70B annual traffic
-
Designed and deployed 2-layer hierarchical model architecture to mitigate a severe class
imbalance problem within a dataset of 10M invoices improving production accuracy from 92%
to 98%; delivered tech talk to team and clients
-
Experimented with 10 useful new XLM features by creating Dataswarm pipelines to extract
features from raw data and enhanced model f1-score by >2%
-
Suggested the cross-functional team with modified category taxonomy based on analyzing
confusion matrix
-
Resolved data leakage by re-splitting train/test datasets so that production accuracy
reflects model generality
Intel, Shanghai, China
Deep Learning Intern, AI and Big Data Dept. | May 2022 - Aug 2022
Project: Spend Classification
-
Built a vertical federated linear regression model that enables companies to collaboratively
train a predictive model of 78% accuracy with 5 GB local data of different feature spaces to
preserve customer privacy
-
Designed and coded a joint computation mechanism of gradient for loss function among several
participants
-
Wrote scripts to encapsulate LibOS and big data applications for training federated learning
(FL) models
-
Evaluated and debugged an LSTM-based horizontal FL model with Intel Software Guard Extensions
NetEase Inc., Hangzhou, China
Android Engineer, Cloud Music Dept. | Jun 2020 - Aug 2020
Project: Revolutionary NetEase Cloud Music Update 8.0
-
Self-learned Kotlin and published detailed study notes to company technical documentation system
- Added new Lottie animation to UI interface as well as beautified VIP profile card display ;
developed new functionalities such as “Follow Anchorman” feature for the radio station in Cloud
Music App
-
Tested and debugged code for robustness; analyzed edge case, usability, and general reliability;
optimized application performance to maintain best practices
- Multidisciplinary Design Program Summer Fellowship, University of Michigan, May 2020
- B. Angell Scholar, University of Michigan, Winter 2021
- Dean’s List, University of Michigan, Fall 2019
- University Honors, University of Michigan, Fall 2019 & Winter 2020
- Most Diversified Team Award for 2020 Tech4Good Hackathon, Ladies Who Tech, December 2020
- Teaching Assistant of CMU DS Master Program Capstone Courses, Carnegie Mellon University, Aug 2022 - Present
- Teaching Assistant of CMU Coding & Algorithms Bootcamp, Carnegie Mellon University, Aug 2022 - Present
- Teaching Assistant of CMU Machine Learning, Carnegie Mellon University, Jan 2022 - May 2022
- Vice President of SJTU-UM Student & Alumni Association, University of Michigan, Dec 2019 -
May 2021
- Instructional Aide of UM EECS 445 Machine Learning, University of Michigan, Jan 2021 - May 2021
- Grader of Advanced Linear Algebra & Machine Learning, University of Michigan, Sept 2020 - Dec 2020
- Transfer Student Leader, University of Michigan, Sept 2020 - May 2021
- 2020 Tech4Good Hackathon, Most Diversified Team Award, Dec 26 2020– Dec 27 2020
- Minister of Brave on Diversity Women Engineers’ Club, UM-SJTU Joint Institute, Feb 2018 - Aug
2019
- Teaching Assistant of German, UM-SJTU Joint Institute, Feb 2019 - May 2019
- Volunteer Teacher of Yuannan Aid Education Program, Yunnan, China, Dec 2018 - Feb 2019
- Volunteer of SJTU Buddy Program, UM-SJTU Joint Institute, Jan 2018 – Dec 2018
- I am an equestrian, I love horsemanship and have been learning it for six years
- I am a blue-belt karateka!
- I paly Chinese zither
- Amateur photographer and singer, I hope you like the ocean picture on the left-hand side :)
- I am also a travel lover, and enjoy wandering around a strange city, my favorite city is Kyoto!