Looking for a job

Yi Zhang

@yizhang2

Research Associate at the University of Augsburg developing generative models for system identification, state estimation, and control.

Germany

Message

What I'm looking for

I’m eager to explore architectures, develop new methods, and translate research ideas into working models, extending my expertise toward world models that connect spatial understanding and physical prediction.

At the University of Augsburg, I develop and validate deep-generative-model-based methods for Bayesian parameter identification, state estimation, dynamics modeling, optimal control, and reinforcement learning under uncertainty.

I’ve applied these methods to industrial problems including MEMS end-of-line testing, steering-system state estimation, anomaly detection, robotics control, and autonomous-driving path planning. I’ve also contributed to collaborative work exploring diffusion models for 3D vehicle generation and vision-language models for CAD generation of mechanical components.

At Trumpf SE + Co. KG, I developed a computer-vision-based surrogate modeling approach for finite-element simulations supporting process-reliable part removal in automated laser cutting. My research interests include generative AI, computer vision, dynamics modeling, and extending this expertise to world models that connect spatial understanding with physical prediction.

Experience

Work history, roles, and key accomplishments

Current

Research Associate

Current

University of Augsburg

Apr 2021 - Present (5 years 6 months)

Developed and validated deep-generative-model-based methods for Bayesian parameter identification, state estimation, dynamics modeling, optimal control, and RL policy under uncertainty. Applied these methods in industrial applications such as MEMS testing, steering system state estimation, anomaly detection, robotics control, and autonomous driving path planning.

Generative Models Bayesian Inference (PyMC)State Estimation Model Predictive Control Reinforcement Learning System Identification Python PyTorch Jax TensorFlow ROS2 Docker Git Linux Slurm

Intern in R&D Pre-development Production Platform MT

Trumpf SE + Co. KG

May 2019 - Oct 2019 (5 months)

Developed a computer-vision-based surrogate modeling approach for finite-element simulations supporting process-reliable part removal in automated laser cutting.

Computer Vision Surrogate Modeling Finite Element Analysis Python

Education

Degrees, certifications, and relevant coursework

University of Augsburg

Doctor of Engineering, Applied Computer Science

2021 -

Doctoral researcher developing probabilistic and generative models for dynamical systems, with a dissertation on deep Bayesian inference for control.

University of Stuttgart

Master of Science, Electrical Engineering

2017 - 2020

Grade: 1.6

Master of Science in Electrical Engineering with a focus on signal processing and machine learning, including a thesis on feature selection using attention mechanisms.