Junsong Wang
创新技术总监 at 云南易见纹语科技有限公司
Haidian District, Beijing, China
Summary
👤
Senior🎓
Top SchoolJunsong Wang is an innovation technology director and former senior research staff member at IBM Research China with over a decade of experience building production AI systems in computer vision, heterogeneous computing, and neural-network hardware acceleration. He has led large-scale projects from IBM PowerAI Vision and airport video analytics to night-driving assistance and a commercial tea-traceability platform, and holds 30+ patents with nearly 20 papers in top conferences including an ICCAD best paper award. Junsong authored the open-source deep neural network compiler AccDNN and developed ultra-low-bit mixed quantization and acceleration architectures (AccELB), demonstrating a rare blend of research, open-source impact, and product delivery. Currently he translates texture-recognition research into deployed anti-counterfeiting and traceability hardware and distributed search systems for agricultural products, notably Pu'er tea. Colleagues describe him as an applied scientist who consistently turns algorithmic innovation into working systems at industrial scale.
9 years of coding experience
硕士, Telecommunications Engineering, 硕士, Telecommunications Engineering at 西安交通大学
Chinese
Github Skills (20)
fpga8
reinforcement-learning8
verilog8
hardware7
machine-learning7
deep-learning7
flexibility7
tensorflow6
accelerator5
semantic-segmentation5
neural-network4
compiler4
object-detection4
caffe3
pytorch3
Programming languages (3)
Github contributions (5)
IBM/AccDNN
Sep 2019 - Dec 2019
A compiler from AI model to RTL (Verilog) accelerator in FPGA hardware with auto design space exploration.
Contributions:18 commits, 10 PRs, 14 pushes in 2 months
rtldesign-space-explorationcompilerdesign-spacehardware
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.