Kairos Power and Oak Ridge National Laboratory use agentic workflows on the Rescale platform for advanced nuclear reactor design and manufacturing

SAN FRANCISCO, Nov. 10, 2026 /PRNewswire/ -- Rescale, the digital engineering platform for the AI era, today introduced new agentic digital engineering capabilities that allow R&D and engineering teams across advanced industries to customize AI agents to their exacting standards, bringing agentic AI into engineering work where every result must be reviewable and traceable.

As engineering organizations continue to deploy more agents into production, the work is shifting from automating individual tasks to running end-to-end engineering processes. That shift requires an engineering-specific agentic harness: the structure and context that govern how agents operate, ensure they meet specific standards, and keep humans in the loop. 

New Agentic Projects on the Rescale platform provide the framework and structure, giving engineers a way to customize agents and build reusable agentic workflows. Teams can define each project with their own context, skills, tools, and guardrails, share projects across teams, and retain approved learnings through project memory. Expanded engineering intelligence capabilities supply the underlying context by connecting data across requirements, design, simulation, test, and process systems into foundational knowledge graphs. As agents execute workflows, the platform generates project-specific work graphs that provide traceability and capture critical learnings. Those learnings enrich agent context, build domain expertise, preserve knowledge, and drive process improvements that make engineering work more efficient and scalable. 

Rescale is also introducing Rescale Desktop, which extends Rescale directly to engineers' local machines, along with expanded Model Context Protocol (MCP) connectors for interoperability across Rescale's leading ecosystem of engineering applications and complementary agentic tools, including Rescale MCP availability on the Claude and Codex directories.

Rescale Edison, the platform's AI assistant and agent orchestrator, now includes persistent memory so engineers can work faster by describing their intent in natural language. Together, these new capabilities work to provide an engineering-specific harness that ensures agents execute to the same standards as experienced engineers, helping organizations develop the trust in agentic outcomes needed to transform operations.

"Our customers build aircraft, vehicles, and reactors, so the agents supporting that work have to deliver reliable results. Agentic Projects provide the needed context that grows with every iteration. This is a material advancement toward an engineering harness built for how these teams actually work," said Jacob Surber, vice president of product at Rescale.

Kairos Power and the Department of Energy's Manufacturing Demonstration Facility (MDF) at Oak Ridge National Laboratory (ORNL), with support from Rescale AI developer engineers, have developed a digital agentic scaffolding on the Rescale platform as part of a collaboration to test advanced manufacturing and construction methods for Hermes 2, Kairos Power's 50-megawatt fluoride-salt-cooled reactor. Working in a shared environment, agents have exercised thermal, flow, and shielding simulation and reporting workflows for the reactor cavity and bioshield plug designs, with engineers approving each agent action. Kairos Power projects that agentic workflows built from past simulations could increase end-to-end simulation analysis efficiency by up to 3x while preserving the data traceability required for regulatory compliance.

"Kairos Power depends on quick learning cycles: design it, build it, test it, and carry what we learn into the next round," said Ed Blandford, chief technology officer and co-founder of Kairos Power. "Much of the difficulty in scaling nuclear energy comes down to manufacturing, from the smallest components to large structures. Rescale brings agentic engineering into the simulations that support that work, connecting our engineers with our ORNL collaborators in one environment built for context and traceability. With agents managing setup, execution, and reporting, our people can spend their time on the engineering judgment and rigor the work demands."

ORNL is encoding its simulation and additive manufacturing expertise in agents on the platform so that engineers can adopt advanced methods to address industry challenges. 

"The Manufacturing Demonstration Facility works because industry engineers and lab researchers solve problems side by side," said Ryan Dehoff, MDF director. "Rescale provides a digital environment for that collaboration. Nuclear innovation demands speed and rigor at the same time, and you can't trade one for the other. Agents allow ORNL's expertise to scale beyond our facility."

The launch also highlights how agentic digital engineering works with Rescale's AI physics solution to make surrogate models easier to build and more accessible to engineering teams. A new dataset creation agent automates the process of turning simulation data into training-ready datasets, reducing the specialized machine learning work involved. This is one of many new AI physics capabilities that help companies put valuable surrogate models into the hands of more engineers:

• Auto Tune and guided pre-processing automate and simplify surrogate model training so organizations can develop better models with fewer specialized resources. 
• Uncertainty quantification builds trust that engineers are using an appropriate surrogate model and flags predictions that should be confirmed with a full-fidelity simulation.
• Design explorer is a new local interface that allows design engineers to use surrogate models to evaluate thousands of design options from their local environment.

For high performance computing (HPC) administrators, new cluster performance metrics show where execution slows and why jobs are waiting. A new priority score considers wait time, job size, and resource usage to help teams allocate resources fairly.

These new and enhanced capabilities across agentic digital engineering, AI physics, and intelligent HPC work together to deliver compounding value for R&D and engineering organizations as they adopt AI-first operations across every stage of their product development lifecycles.

Organizations can learn more about the new capabilities at rescale.com.

About Rescale

Rescale is the digital engineering platform for the AI era. The R&D and engineering teams shaping the future across aerospace, automotive, defense, energy, public sector, life sciences, and manufacturing trust the Rescale platform to accelerate innovation. Rescale integrates agentic digital engineering, AI physics, advanced modeling and simulation, and intelligent HPC solutions on a single platform to unify product data, automate workflows, optimize resources, and bring breakthroughs to market faster.

Rescale is backed by leading investors such as NVIDIA, Sam Altman, Jeff Bezos, Paul Graham, Peter Thiel, Microsoft, Hitachi, and University of Michigan.

For more information, visit rescale.com.

View original content to download multimedia:https://www.prnewswire.com/news-releases/rescale-expands-agentic-digital-engineering-with-customizable-ai-agents-for-rd-teams-302934651.html

SOURCE Rescale Inc.