Max Robotics — Global Service Robot Directory + US Certification

About

Tinoosh Mohsenin is a prominent researcher at the intersection of artificial intelligence hardware optimization, efficient deep learning, and autonomous systems. Her work spans two compelling frontiers: the design of neural network accelerators for edge and embedded AI, and the development of hierarchical reinforcement learning frameworks for real-world robotics and autonomous agents. Mohsenin's most influential contribution is her comprehensive survey on optimizing neural network accelerators for micro-AI on-device inference (2021, 63 citations), which has become a key reference for researchers and engineers working to deploy deep neural networks efficiently on resource-constrained hardware. This work addresses the critical challenge of balancing inference accuracy with computational efficiency — a cornerstone problem in the TinyML and edge AI communities. Her more recent research on vision transformer fine-tuning for TinyML platforms reflects her continued leadership in hardware-aware AI deployment. Equally notable is her pioneering work in hierarchical reinforcement learning, where she explores how language models and semantic goals can help agents tackle long-horizon tasks more effectively. Through projects like ReProHRL and LLM-augmented hierarchical agents, Mohsenin bridges theoretical AI research with practical robotic applications. Her diverse and growing body of work positions her as an innovative voice shaping the future of intelligent, efficient, and autonomous systems.