Reazon Holdings
Research Scientist - AI/Foundation models
undefined
Tags: Full-time, 6~8 YOE, Remote
Remote (Remote)・Fetched 5 days ago
Job Description
About Enactic
"Bringing autonomous humanoid companions into homes worldwide to assist and enhance our daily lives"Enactic is a deeptech startup on a mission to ship assistive humanoid robots that enhance our independence and quality of life in a world facing critical labor shortages, due to aging demographics.
Responsibilities
As our AI and Foundation Model Reseach Scientist, you will build Physical AI systems for humanoid robots that assist people in home environments. Your mission is to establish a continuous improvement cycle between real-world data collection and perception-decision-action models, creating AI that's safe, practical, adaptive, dexterous, and robust.
Requirements
・Experience building AI systems that perform in real-world environments through experimentation with world models or foundation models・Ability to lead the design and validation of continuous improvement cycles: real-world data → model → action → feedback・Experience developing and operating production-quality libraries and systems in C/C++, Python, Docker, Linux, and Git
Bonus Qualifications
・Experience implementing force control, compliance control, and bilateral control in real-world environments・Experience creating or making significant contributions to widely-used open-source projects・Experience designing large-scale multimodal data pipelines for acquisition, synchronization, and preprocessing・Experience designing algorithms for behavior models, such as stable learning, safe exploration, and online adaptive control・Experience bridging technology development and social acceptance through design and real-world implementation・Experience with distributed training, optimization, and hyperparameter tuning of billion-parameter-scale deep learning models・Understanding of theory and implementation for cutting-edge methods including representation learning, world models, imitation learning, and reinforcement learning・Deep understanding of humanoid mechanics, actuator control, and sensor integration・Experience developing systems applying deep learning to whole-body motion control, locomotion, and manipulation・Strong communication skills to drive cross-disciplinary collaboration
"Bringing autonomous humanoid companions into homes worldwide to assist and enhance our daily lives"Enactic is a deeptech startup on a mission to ship assistive humanoid robots that enhance our independence and quality of life in a world facing critical labor shortages, due to aging demographics.
Responsibilities
As our AI and Foundation Model Reseach Scientist, you will build Physical AI systems for humanoid robots that assist people in home environments. Your mission is to establish a continuous improvement cycle between real-world data collection and perception-decision-action models, creating AI that's safe, practical, adaptive, dexterous, and robust.
Requirements
・Experience building AI systems that perform in real-world environments through experimentation with world models or foundation models・Ability to lead the design and validation of continuous improvement cycles: real-world data → model → action → feedback・Experience developing and operating production-quality libraries and systems in C/C++, Python, Docker, Linux, and Git
Bonus Qualifications
・Experience implementing force control, compliance control, and bilateral control in real-world environments・Experience creating or making significant contributions to widely-used open-source projects・Experience designing large-scale multimodal data pipelines for acquisition, synchronization, and preprocessing・Experience designing algorithms for behavior models, such as stable learning, safe exploration, and online adaptive control・Experience bridging technology development and social acceptance through design and real-world implementation・Experience with distributed training, optimization, and hyperparameter tuning of billion-parameter-scale deep learning models・Understanding of theory and implementation for cutting-edge methods including representation learning, world models, imitation learning, and reinforcement learning・Deep understanding of humanoid mechanics, actuator control, and sensor integration・Experience developing systems applying deep learning to whole-body motion control, locomotion, and manipulation・Strong communication skills to drive cross-disciplinary collaboration