Yijia Liu
Algorithm Engineer at ByteDance
Hangzhou City, Zhejiang, Singapore
Summary
🤩
Rockstar🎓
Top SchoolYijia Liu is an algorithm engineer with 13 years of experience building and productionizing NLP and ML systems across leading tech companies including ByteDance and Alibaba. Holding a Ph.D. from Harbin Institute of Technology and a visiting stint at University of Washington, he blends deep academic research with hands-on engineering to deliver robust language models and tooling. His open-source contributions include cross-lingual ELMo pretraining work and practical enhancements to DyNet and pyltp, reflecting expertise in model internals, multilingual embeddings, and Python-C++ integration. At Alibaba he served as an Algorithm Expert and now applies that expertise at ByteDance to scale algorithms in real-world products. Colleagues describe him as the kind of engineer who can both refactor low-level C++ for Windows compatibility and extend deep learning pipelines to support layer-specific outputs—a combination of pragmatism and research-driven rigor. Based in Hangzhou with ties to Singapore, he brings a global perspective to large-scale language engineering challenges.
13 years of coding experience
1 year of employment as a software developer
visiting student, Computer Science, visiting student, Computer Science at University of WashingtonPh.D., Computer Science, Ph.D., Computer Science at Harbin Institute of Technology
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (19)
pytorch10
c-language10
dynet10
python10
machine-learning10
deep-learning10
natural-language-processing10
neural-network10
nlp10
cprogramming-language10
nlpjs9
resource-loading9
data-loading9
preloading9
boost-python9
Programming languages (6)
Github contributions (5)
HIT-SCIR/ELMoForManyLangs
Jul 2018 - May 2020
Pre-trained ELMo Representations for Many Languages
Role in this project:
ML Engineer
ML EngineerContributions:24 commits, 4 PRs, 20 pushes in 1 year 10 months
Contributions summary:Yijia contributed to the development and maintenance of pre-trained ELMo representations for multiple languages. Their work focused on adding features such as dumping specific layers of the ELMo model and supporting text output formats. The changes included modifications to the `gen_elmo.py` file to incorporate these features and adjustments to the `elmo.py` file related to model loading and embedding extraction. Additional changes involved merging branches and refactoring code to improve programmatically API.
nlprepresentationsbertpre-trainedmachine-learning
HuangFJ/pyltp
Feb 2014 - Nov 2017
LTP wrapper module for Python. Move to https://github.com/HIT-SCIR/pyltp
Role in this project:Back-end Developer
Back-end DeveloperContributions:76 commits in 3 years 10 months
Contributions summary:Yijia primarily contributed to the development of a Python extension for the Language Technology Platform (LTP). Their work focused on adapting the code to LTP 3.0, adding example usages, and supporting new functionalities like parsing and named entity recognition. They also introduced sentence splitting capabilities and refactored code to align with updated LTP APIs. The user demonstrated a clear understanding of the LTP library and its integration with Python.
hitpythonpython3
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.