AI-powered humanoid robot platform with mentorship-based learning
Mentee Robotics builds a hardware-software stack for general-purpose humanoid robots centered on learnable AI models. The tech foundation spans low-level robotics (ROS 2, EtherCAT, CAN) through high-level ML (PyTorch, TensorFlow, CUDA) to cloud orchestration (AWS, Kubernetes), indicating a fully integrated approach from simulation to edge deployment. Active projects on foundation models, memory systems, and real-time autonomy suggest the team is solving two core constraints: scaling deep learning for robot datasets and optimizing inference on edge hardware—both listed as documented pain points.
Mentee Robotics, founded in 2022 and based in Herzliya, Israel, develops a general-purpose humanoid robot platform designed around an AI-based mentorship model. The company operates a 51–200-person team heavily weighted toward engineering and research, with all hiring currently domestic. The platform integrates simulation pipelines, perception and autonomy stacks, and a robot memory system, deployed across both cloud infrastructure (AWS, Kubernetes) and embedded Linux devices. Current product development focuses on scaling deep learning frameworks for robot training datasets and optimizing inference for real-time performance on constrained hardware.
Python, C++, PyTorch, TensorFlow, ROS 2, OpenCV, CUDA, AWS, Kubernetes, Docker, plus embedded protocols (EtherCAT, CAN, RS485) and design tools (Altium Designer, OrCAD).
Deep learning infrastructure, simulation training pipelines, edge deployment optimization, foundation models for humanoids, real-time autonomy stacks, and a robot memory database.
Other companies in the same industry, closest in size