#USA #San Francisco #Digital Engineering #AI Agents #Rescale

Rescale's Breakthrough in Digital Engineering

In a significant stride toward enhancing efficiency in research and development (R&D), Rescale has launched a new suite of customizable agentic digital engineering capabilities. This innovative approach is designed to empower R&D and engineering teams across various advanced industries, facilitating the adaptation of AI agents to their precise needs—resulting in workflows that are not only reliable but also fully traceable.

With the growing need for teams to deploy more agents into production environments, the focus is shifting from merely automating singular tasks to managing comprehensive engineering processes end-to-end. This pivotal change necessitates a tailored engineering-specific framework that contextualizes how agents function, ensuring that results adhere to established standards while maintaining human oversight.

Introducing Agentic Projects

The newly introduced Agentic Projects feature on the Rescale platform provides a robust structure that enables engineers to customize agents and create reusable workflows. This platform allows teams to define unique project contexts, skillsets, tools, and boundaries, facilitating knowledge sharing and learning retention through a project memory feature. By connecting data from requirements, design, simulation, testing, and various process systems into foundational knowledge graphs, Rescale enhances the contextual understanding of its agents.

Each time agents execute workflows, the platform generates project-specific work graphs. These not only provide traceability but also capture essential lessons that can be used to enrich the agents' context and inform future processes. This capability is poised to enhance domain expertise and drive further improvements in engineering efficiency and scalability.

New Tools and Features

Alongside these advancements, Rescale has also announced the launch of Rescale Desktop, which allows engineers direct access to the platform from their local machines. Additionally, enhanced Model Context Protocol (MCP) connectors improve interoperability across Rescale’s ecosystem of engineering applications, including resources like Claude and Codex directories.

Rescale Edison, the platform's AI assistant, now features persistent memory. This upgrade allows engineers to relay their intentions in natural language, significantly streamlining their workflow and accelerating project timelines. Together, these innovations create a unique engineering harness that ensures the standards upheld by expert engineers are mirrored in the agents’ outcomes.

Industry Applications and Collaborations

This transition to agentic workflows is already being pursued by leading organizations. Notably, Kairos Power and the Oak Ridge National Laboratory's (ORNL) Manufacturing Demonstration Facility (MDF) have already adopted these capabilities on the Rescale platform. Their collaboration focuses on testing advanced manufacturing methods for a fluoride-salt-cooled reactor, known as Hermes 2. Through this partnership, agents have facilitated thermal, flow, and shielding simulation workflows, with engineers verifying each action taken by the agents.

Kairos Power anticipates that these newly integrated workflows could enhance the efficiency of simulation analysis by up to threefold while ensuring essential data traceability for regulatory compliance.

Ed Blandford, the company's Chief Technology Officer, emphasizes the importance of rapid learning cycles through the design, build, test, and learn methodology. He acknowledges the challenges in scaling nuclear capabilities due to manufacturing complexities, highlighting the critical role Rescale plays in bridging these gaps, especially in simulations tied to engineering judgments that require meticulous attention.

ORNL is poised to leverage its simulation and additive manufacturing expertise by embedding methodologies within agents. This collaboration ensures that both industry engineers and lab researchers effectively tackle challenges collectively, affirmed by Ryan Dehoff, MDF Director, who notes that nuclear innovation demands balancing speed with meticulousness, an expectation that agentic processing meets aptly.

Enhancing Surrogate Modeling

Moreover, Rescale's integration of agentic digital engineering with AI physics innovations is noteworthy. A new agent dedicated to dataset creation simplifies the transformation of simulation output into training-ready datasets. This simplifies the complexities surrounding machine learning and fosters broader accessibility of surrogate models to engineering teams.

Improvements in training automation like Auto Tune streamline surrogate model training, allowing organizations to enhance model quality with fewer specialized resources. Additionally, new uncertainty quantification tools help validate model appropriateness and spot areas requiring high-fidelity simulations.

For HPC administrators, new performance metrics illuminate the causes behind execution delays, serving as a resource allocation framework that provides fair access to essential computing resources.

Conclusion

As R&D and engineering organizations embrace AI-first strategies throughout their product development cycles, these newly introduced capabilities from Rescale are set to significantly improve their operational effectiveness. For further details about these innovations, visit rescale.com.

About Rescale

Rescale stands as a pioneering digital engineering platform ushering in a new era characterized by AI-driven innovations. By unifying agentic digital engineering and advanced simulation capabilities, it serves industries like aerospace, automotive, defense, energy, and many more—all striving to accelerate their paths to innovation.