Dejian Zhou
Undergraduate Student Researcher
Beijing, Switzerland
Summary
🤩
Rockstar🎓
Top SchoolDejian Zhou is an undergraduate mechanical engineering student at ETH Zürich with four years of hands-on experience building AI systems and startups across Europe. Currently balancing roles as an AI Engineer at Verdant Rock and an Undergraduate Student Researcher at Robotic Systems Lab, he focuses on practical ML problems such as out-of-distribution detection—contributing evaluator improvements to the OpenOOD benchmark. He co-founded a venture (PVelion) and has driven logistics and entrepreneurship initiatives at ETH, demonstrating an unusual mix of research, product and operational experience for his level. Past internships in innovation transfer and finance-club leadership reflect strong communication and project-management skills alongside technical depth. Based in Beijing with ties to Switzerland, he brings cross-cultural adaptability and a track record of turning research ideas into demonstrable prototypes.
4 years of coding experience
1 year of employment as a software developer
Edvard Munch High SchoolMassachusetts Institute of TechnologyGerman High School Diploma, German High School Diploma at Gymnasium Dresden-BühlauBachelor, Mechanical Engineering, Bachelor, Mechanical Engineering at ETH Zürich
Github Skills (5)
pytorch10
machine-learning10
anomaly-detection10
python10
cosine-similarity9
Programming languages (2)
Github contributions (5)
Jingkang50/OpenOOD
Feb 2022 - Aug 2022
Benchmarking Generalized Out-of-Distribution Detection
Role in this project:
ML Engineer
ML EngineerContributions:5 reviews, 45 commits, 10 PRs in 6 months
Contributions summary:Dejian primarily contributed to the `openood/evaluators/kdad_evaluator.py` file, focusing on the implementation and refinement of an anomaly detection evaluator. The code changes involved modifications to the evaluation logic, including the incorporation of different loss functions and similarity calculations, suggesting an effort to improve the model's OOD detection capabilities. Further modifications include adjusting the dataset used in the testing. The work points towards an active involvement in model evaluation and performance optimization within the context of out-of-distribution detection.
benchmarkingoutlier-detectionrobustnessanomaly-detectionbenchmark
JediWarriorZou/OpenOOD
Jun 2022 - Jan 2023
Benchmarking Generalized Out-of-Distribution Detection
Contributions:13 pushes in 6 months
benchmarkingbenchmarkout-of-distribution-detectionout-of-distributiondistribution
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.