Robotics Software Engineer, Robot Learning and Manipulation, Singapore

Role
As a Robotics Software Engineer specializing in Robot Learning and Manipulation, you will drive the design, training, validation and deployment of contact-rich manipulation skills for real-world manufacturing. Your focus will be on solving dexterous, contact-rich assembly tasks. You will develop and integrate advanced sensor-based control strategies and Machine Learning (ML) policies within the Intrinsic Robot Control stack, managing the full model lifecycle - from architecture, large-scale data collection and training to validation, deployment and cycle-time optimization - to meet rigorous industrial standards. How your work moves the mission forward
Develop, train, and deploy multi-modal feedback- and interaction controllers to solve high-precision insertion and assembly tasks. Lead the AI manipulation model lifecycle, conducting regular trials on industrial hardware to evaluate algorithmic changes and curate high-quality training datasets. Architect and build modular components for reinforcement learning and imitation learning, continuously seeking to improve the robustness of contact-rich assembly tasks. Stress-test and optimize the real-time execution framework for learned manipulation models within the Intrinsic platform. Skills you will need to be successful
PhD or equivalent professional experience in Robot Learning and AI Manipulation (e.g., Applied RL, Visuomotor Policy Learning, Vision-Language-Action models). 2 years of professional experience in
C++ and Python , with a proven track record of shipping production-quality code. ML Frameworks: Deep expertise with JAX, TensorFlow, or PyTorch. Hardware Experience: Direct experience testing and iterating on physical robots integrated with vision and force-torque sensors. Operational Grit: A desire to spend significant time in the lab, managing hardware experiments and troubleshooting physical-world edge cases. Communication: Business fluency in English. Skills that will differentiate your candidacy
Multimodal Perception and Control: Expertise in integrating tactile/force sensing and computer vision for contact-rich tasks. Hardware Experience: Direct experience testing and iterating on physical robot manipulators integrated with vision and force-torque sensors. Large-Scale Training: Proficiency in training on massive multimodal datasets and managing cloud-based training workflows (e.g. Argo). Infrastructure :
Experience with cloud-based training (Argo) and deploying inference on the edge. Customer Focus: Willingness to travel internationally to analyze customer use cases and deploy solutions under ambitious timelines.