3D AI technology startup Endotlight announced on the 29th that it signed a memorandum of understanding (MOU) with AI data construction specialist Crowdworks for the joint development of a 3D manufacturing AI data pipeline solution.
This agreement was launched with the aim of addressing issues that have repeatedly arisen in the manufacturing, robotics, defense, and physical AI sectors, such as data shortages, high deployment costs, and the absence of integrated management systems.
The two companies plan to jointly develop an AI data pipeline solution that can be utilized across industries by combining Endotrite's AI-based 3D CAD and synthetic data generation technology 'TRINIX', Crowdworks' data construction solution 'Workstage', and high-quality LLM and vision data processing capabilities.
Through this, we aim to build an end-to-end data chain from data creation to labeling, quality verification, and delivery, and provide customers with a complete dataset rather than simply raw data.
In particular, this collaboration will also promote the development of a next-generation robot learning dataset that integrates CAD-based synthetic data with multimodal information, including teleoperation behavioral data, sensor feedback, and video and audio. This integrated approach, encompassing "behavior + environment + sensor + language," is expected to have significant applications in robotics and physical AI.
The two companies plan to strengthen their market position through joint sales and marketing efforts, and to collaborate on new business opportunities, including joint research and development for data standardization and quality improvement, technology consulting, and service development in the manufacturing, defense, and robotics sectors.
“Securing high-quality robot and physical AI data is one of the key challenges facing the industry today,” said Jinyoung Park, CEO of Endotlight. “By combining CAD simulation-based Trinix data with behavioral and sensor-based multimodal data, we will be able to build a globally competitive full-stack robot learning dataset.”
- See more related articles
You must be logged in to post a comment.