Rapid advances in AI are transforming engineering design, reshaping not only how products are developed but also how design decisions are made. As illustrated above, this evolution spans three paradigms: Optimal Design, in which engineers use simulation-based optimization and CAD/CAM/CAE tools to improve product performance; Generative Design, in which generative AI expands design exploration while humans remain the primary decision-makers; and Autonomous Design, in which AI takes a leading role in formulating problems, selecting tools, and developing solutions under human oversight. This progression reflects a fundamental shift from AI as a tool to AI as a decision-maker.

Realizing Autonomous Design requires AI systems that can both reason about physical behavior and act purposefully throughout the design process. Physics AI grounds design exploration and evaluation in physical principles, enabling AI to predict and assess how products will perform. Agentic AI enables systems to plan workflows, coordinate engineering tools, and iteratively refine solutions. Together, these capabilities provide the foundation for AI to carry out increasingly complex design processes with greater autonomy.

At Smart Design Lab, we bring Physics AI and Agentic AI together to advance Autonomous Design and drive the AI transformation of product development in the manufacturing industry. By connecting advances in AI with practical engineering workflows, we aim to accelerate development, improve product performance, and enable new design possibilities.

Our ultimate mission is to democratize design with AI, empowering everyone to create the best possible products.

The framework shown here connects three complementary capabilities:

① Generative AI: Generating a Large Number of Aesthetic and Engineering-Feasible Design Candidates  
  - Generative Design + AI
  - Design + Agentic AI 
  - Text-to-CAD
  - Topology Optimization + AI
  - Parametric Design + AI
  - Design Exploration + AI  

② Predictive AI: Predicting the Engineering Performance, Manufacturability, Aesthetics, and Other Factors in Real-Time
  - Engineering Performance Prediction + AI
  - Manufacturability Prediction + AI
  - Customer Preference (Aesthetics) Prediction + AI

③ Optimization AI: Optimizing Design Solutions in Real-Time to Achieve Target Performance and Satisfy Constraints 
  - Inverse Design + AI
  - Design Optimization + AI
  - Multidisciplinary Design Optimization + AI
  - Design Recommendation + AI