Xichen Pan
Visiting Researcher at Meta
New York, New York, United States
Summary
🤩
Rockstar🎓
Top SchoolXichen Pan is a PhD candidate in Computer Science at NYU Courant and a Visiting Researcher at Meta AI, specializing in controllable generative models and vision-centric multimodal systems. With seven years of research and engineering experience across Meta, Microsoft Research Asia, Alibaba, and Horizon Robotics, he blends deep academic rigor with hands-on contributions to large open-source multimodal projects like Microsoft's unilm. His work spans model development, image-processing integrations (e.g., Kosmos-G controlnet updates), and full-stack tooling to streamline research-to-inference pipelines. A Shanghai Jiao Tong University alumnus with a Best Thesis Award, he brings a pragmatic focus on deployable research and a knack for bridging cutting-edge models with production-ready code.
7 years of coding experience
Doctor of Philosophy - PhD, Computer Science, Doctor of Philosophy - PhD, Computer Science at New York UniversityBachelor of Engineering - BE, Computer Science, Bachelor of Engineering - BE, Computer Science at 上海交通大学
Github Skills (19)
python10
pre-trained-model10
image-processing10
net10
css10
multimodal10
llm10
computer-vision10
nlp10
computer-science10
application10
mkdocs9
pytorch9
yaml9
html8
Programming languages (3)
Github contributions (5)
opencsapp/opencsapp.github.io
Mar 2022 - Jan 2023
Open CS Application | 开源CS申请
Role in this project:
Full-stack Developer
Full-stack DeveloperContributions:493 commits, 173 PRs, 915 pushes in 9 months
Contributions summary:Xichen primarily worked on enhancing the front-end of the application and integrating the Giscus comment system. They created a script using Python to dynamically generate content for `mkdocs.yml` and `docs/选校梯度.md`, indicating back-end and build process involvement. The user also made several CSS updates to modify the site's appearance, indicating UI/UX focus, along with modifying the main HTML file.
mastersmasterapplicationcomputer-sciencegraduate-application
microsoft/unilm
Mar 2023 - Mar 2024
Large-scale Self-supervised Pre-training Across Tasks, Languages, and Modalities
Role in this project:ML Engineer
ML EngineerContributions:7 PRs, 8 comments in 1 year
Contributions summary:Xichen's contributions focused on the development and modification of a multimodal model, specifically for image-related tasks. They initialized and updated code for the "kosmos-g" component, which likely involved integrating image processing functionalities. The commits show modifications to controlnet applications for tasks like canny edge detection and other visual transformations, as well as configuration updates, and file conversions. They also updated steps for the inference process.
layoutxlmtraininglanguage-understandingvision-and-languagewavlm
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.