Jiawang Bian - Research Scientist at Bytedance

Jiawang Bian

Research Scientist at Bytedance

Greater Adelaide Area Australia

Join Prog.AI to see contacts

Summary

👤

Senior

🎓

Top School

Jiawang Bian is a research scientist specializing in computer vision with over five years focused on vision research and 11 years of engineering experience overall. Based in Adelaide, he has contributed to industry-leading labs via internships at Facebook, AWS, TuSimple and research roles across SUTD and ADSC while pursuing a PhD at the University of Adelaide. His open-source work includes practical C++ and MATLAB improvements to the widely used GMS feature matcher and engineering of training/evaluation pipelines for scale-consistent unsupervised depth learning (SC-SfMLearner). He combines hands-on systems tuning—thresholds, scaling, rotation handling and loss tweaks—with reproducible ML workflows, bridging algorithmic research and production-ready implementations. Colleagues would note his knack for refactoring legacy code into maintainable, performant modules that accelerate experimentation.

11 years of coding experience

2 years of employment as a software developer

Bachelor of Engineering - BE, Computer Science, Bachelor of Engineering - BE, Computer Science at 南开大学

Doctor of Philosophy - PhD, Computer Science, Doctor of Philosophy - PhD, Computer Science at University of Adelaide

Github Skills (22)

depth-estimation10

pytorch10

c-language10

python10

machine-learning10

unsupervised-learning10

computer-vision10

cprogramming-language10

preprocess9

preprocessing9

algorithm9

code-optimization9

algorithms9

opencv9

optimisation9

Programming languages (3)

C++Jupyter NotebookPython

Github contributions (5)

JiawangBian/GMS-Feature-Matcher

Oct 2016 - Jun 2020

C++ code for "GMS: Grid-based Motion Statistics for Fast, Ultra-robust Feature Correspondence"

Role in this project:

Back-end Developer

Contributions:93 commits, 4 PRs, 98 pushes in 3 years 8 months

Contributions summary:Jiawang's contributions center around modifying and improving the GMS (Grid-based Motion Statistics) feature matching algorithm. Their work involves adjusting thresholds, updating demo implementations, and rewriting and refactoring the GMS matcher code. They appear to be focused on optimizing the matching process, as evidenced by changes to parameters and the introduction of scaling and rotation options. This includes modifications to C++ and MATLAB code.

statisticshomographymotionultracorrespondence

JiawangBian/SC-SfMLearner-Release

Aug 2019 - Sep 2022

Unsupervised Scale-consistent Depth Learning from Video (IJCV2021 & NeurIPS 2019)

Role in this project:

ML Engineer

Contributions:1 release, 53 commits, 47 pushes in 3 years

Contributions summary:Jiawang primarily contributes to the training and evaluation pipeline for unsupervised depth and ego-motion learning from monocular video, as indicated by edits to the `train.py`, `loss_functions.py`, and testing scripts. They modified the parameter parsing in `train.py`, adjusted the loss functions to include ssim and mask, and updated image resizing functionality in `test_vo.py` and `test_pose.py`. The commits also involve updating and adding scripts, like `test_kitti_depth.sh`, showing the user is actively working to improve model training and evaluation processes.

pytorchvisual-odometrykittideep-learningunsupervised

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.

Request Free Trial