Michelle Valente

Product manager for Calvin at Wandercraft since September 2025, responsible for the headless industrial humanoid the company built for factory work. Her route into the role is unusual for product: she is a robotics PhD whose research was about the precise problem a legged robot on a plant floor has to solve, which is knowing where it is.

Her doctorate at Mines Paris (PSL), defended on 18 December 2019 at the Centre for Robotics under Arnaud de La Fortelle, was titled "SLAM and data fusion for autonomous vehicles: from classical approaches to deep learning methods". The through line of the papers she published with Cyril Joly and de La Fortelle is sensor fusion under uncertainty. Two 2018 ICARCV papers fused a 2D laser scanner with a stereo camera in evidential grid maps, built on Dempster-Shafer theory rather than a single probability, and then localized by matching those grids; a 2019 journal version followed in Unmanned Systems. Her IROS 2019 paper, "Deep Sensor Fusion for Real-Time Odometry Estimation", presented what it describes as the first framework to fuse a 2D laser scanner and a mono camera for odometry using convolutional networks, notably without requiring the two sensors to be calibrated against each other, recasting the pose estimate as an ordinal classification problem and evaluating on the KITTI benchmark.

That research lineage sits directly under Calvin-40. Calvin has no head: it perceives through a waist mounted RGBD camera, inertial measurement units in each limb and force sensors under its feet, and it has to keep its balance and know its position while carrying loads its makers put at 40 to 50 kg through a working factory. Fusing sensors that disagree into one usable estimate of where a machine is, is the problem she spent her doctorate on.

Between the PhD and Wandercraft she worked the applied side of the same stack. She was a robotics engineer at Stanley Robotics from 2020 to 2021, working on perception and localization for its autonomous car parking robots. She then spent three years at MathWorks as an application engineer for computer vision and robotics, supporting customers across automotive, medical devices, mobile robots, drones and robot arms. In 2025 she was technical product manager for robotics and AI at Stereolabs, the French depth camera maker, leading product strategy for hardware and software across off road and humanoid applications, before moving to Wandercraft the same year. She holds an engineering degree in computer engineering from Pontificia Universidade Catolica do Rio de Janeiro.