Max Robotics — Global Service Robot Directory + US Certification

About

Michiel Hermans is a leading researcher at the intersection of robotics, machine learning, and control theory, best known for pioneering work in differentiable physics engines. His most influential contribution, the 2019 paper "A Differentiable Physics Engine for Deep Learning in Robotics" (176 citations), fundamentally challenged the field's reliance on treating robots as black boxes during controller optimization. By enabling gradient-based optimization through physical simulations, Hermans opened the door for more efficient, data-driven approaches to robotic control—moving beyond traditional evolutionary algorithms and reinforcement learning. His broader research explores the automated design of complex dynamic systems, where computation and physical design converge to create robots requiring minimal control effort. Hermans has also contributed to modular control architectures inspired by biological motor primitives, advancing our understanding of how complex movements like walking and reaching can be decomposed and learned. Beyond technical contributions, he has championed robotics education, using robot building as a motivational tool to engage students in engineering. With a career spanning foundational theory to practical implementation, Hermans continues to shape how robots learn, move, and are designed.