Yanyi Liu
Business Operations
Haidian District, Beijing, China
Summary
🤩
Rockstar🎓
Top SchoolYanyi Liu is a seasoned operations professional with 11 years of experience based in Haidian District, Beijing, combining business operations at IBM with a strong hands-on background in machine learning engineering from active open-source contributions. Though currently titled in Business Operations, Yanyi has made substantive technical contributions to prominent ML projects—helping integrate datasets and model configs in OpenMMLab’s mmrotate and adding model-parallel support for Qwen2 variants in the high-throughput vLLM inference engine. This blend of operational rigor and technical depth enables them to bridge engineering teams and product execution, improving reproducibility and deployment readiness. A Northeastern University MS alum with earlier editorial and executive-assistant roles, Yanyi brings cross-functional communication skills and an eye for detail that surface in both project configuration work and operational processes.
11 years of coding experience
1 year of employment as a software developer
Bachelor's degree, Bachelor's degree at Central University of Finance and EconomicsMaster of Science, Master of Science at Northeastern University
English, Chinese, Chinese
Github Skills (19)
pytorch10
system-configuration10
parallelization10
python10
inference10
llm10
custom-configuration10
object-detection10
computer-vision10
transformer10
configurations10
yml-configuration10
data-set9
datasets9
mlops9
Programming languages (6)
Github contributions (5)
open-mmlab/mmrotate
Mar 2022 - Dec 2022
OpenMMLab Rotated Object Detection Toolbox and Benchmark
Role in this project:
Backend Developer
Backend DeveloperContributions:81 reviews, 14 commits, 26 PRs in 9 months
Contributions summary:Yanyi's commits primarily involve adding initialization configurations (`init_cfg`) to various head components and modifying existing ones, such as `KFIoUODMRefineHead` and `ODMRefineHead`. They also addressed Windows compatibility issues, including switching to original `pycocotools` and removing `opencv`. Furthermore, the user contributed to the integration of the HRSC dataset, suggesting a focus on dataset integration and model configuration within the framework. The changes indicate involvement in model definition and configuration aspects of the project.
pytorchdetectiontoolboxefficientnetobject-detection
vllm-project/vllm
Aug 2023 - Mar 2025
A high-throughput and memory-efficient inference and serving engine for LLMs
Role in this project:ML Engineer
ML EngineerContributions:10 reviews, 9 PRs, 26 comments in 1 year 7 months
Contributions summary:Yanyi's commits primarily focus on supporting new models and features within the vLLM project, particularly those related to model parallelism (PP) for various Qwen2 models. They added support for Qwen2-VL-2B-Instruct, modifying code to incorporate PP, integrating new models and configurations. They also contributed to the codebase by fixing tool parser errors, improving error handling in the OpenAI serving chat, and adding the ability to load generation config from the model.
amdcudadeepseekgpthpu
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.