Haotian Liu
Software Engineer at TikTok
California, United States
Summary
🤩
Rockstar🎓
Top SchoolHaotian Liu is a software engineering leader with 11 years of experience building large-scale ads, growth integrity, and cloud infrastructure systems across TikTok, Google, Uber, and Quantcast. Currently leading Account & User Growth Integrity at TikTok USDS and previously driving Ads Platform engineering, he combines product-focused leadership with deep hands-on experience shipping reliable, high-throughput services. His background includes cloud directory work at Google and mobile infrastructure at Uber, giving him strong system design and operational instincts. An MS in Computer Engineering from Northwestern complements his practical expertise in production ML/infra; he has contributed to real-time edge instance segmentation tooling such as optimizing YOLACT Edge for TensorRT and FP16. Known for improving evaluation pipelines and integrating performance-focused configurations, he brings a pragmatic approach to bridging ML research and production constraints. Based in California, he pairs technical depth with cross-functional collaboration to turn complex requirements into dependable, scalable systems.
11 years of coding experience
9 years of employment as a software developer
Bachelor of Engineering, Technology &Apparatus of Measuring and Control, Bachelor of Engineering, Technology &Apparatus of Measuring and Control at Beihang UniversityMaster of Science, Computer Engineering, Master of Science, Computer Engineering at Northwestern University
Stackoverflow
886reputation
24kreached
34answers
4questions
Github Skills (15)
tensorrt10
pytorch10
computer-vision10
instance-segmentation10
configuration-management9
realtime8
json6
htaccess6
html6
wordpress6
foreach6
php6
apache6
forloop6
mysql6
Programming languages (9)
Github contributions (5)
WisconsinAIVision/yolact_edge
Dec 2020 - Dec 2022
The first competitive instance segmentation approach that runs on small edge devices at real-time speeds.
Role in this project:
ML Engineer
ML EngineerContributions:4 reviews, 71 commits, 16 PRs in 1 year 11 months
Contributions summary:Haotian primarily focused on modifying configuration files and code related to the YOLACT edge instance segmentation model. Their contributions included fixing configurations for different training scenarios (e.g., trainflow, without TensorRT, FP16), integrating new configurations, and cleaning up TensorRT-related code. Further, the user worked on addressing an issue related to model evaluation and improving the TensorRT integration.
pytorchrealtimecompetitivesegmentationdeep-learning
haotian-liu/LLaVA
Apr 2023 - Apr 2024
[NeurIPS'23 Oral] Visual Instruction Tuning (LLaVA) built towards GPT-4V level capabilities and beyond.
Contributions:6 releases, 8 reviews, 66 PRs in 1 year
chatbotchatgptfoundation-modelsgpt-4instruction-tuning
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.