The problem

Wearable design involves complex tradeoffs between form, fit, materials, and user constraints. Exploring the design space manually is slow and leaves good solutions undiscovered. Designers need assistance that understands 3D geometry, user context, and design goals—without replacing human judgment.

The solution

AIDC is a multi-agent AI system that combines generative AI, 3D geometric processing, and predictive ML models to support autonomous design assistance for Samsung's wearable technology. Agents collaborate on design exploration, feasibility, and optimization while keeping the designer in the loop.

Without AIDC

Manual exploration of design space, slow iteration, and limited use of historical design data and simulation.

With AIDC

Multi-agent orchestration for design suggestions, 3D body modeling, and predictive evaluation—faster exploration with human oversight.

What it does

  • Multi-agent architecture – LangChain/LangGraph orchestration for specialized design, geometry, and evaluation agents.
  • 3D body modeling – Geometric processing for fit and form constraints.
  • Generative AI – Design suggestions and variations grounded in context.
  • Predictive ML – Models for feasibility and performance estimation.
  • Production deployment – FastAPI, FAISS for retrieval, integrated into design workflows.

Tech stack

LangChain, LangGraph, PyTorch, FAISS, FastAPI. Built for accuracy and integration with existing design tools.

Next steps

Roadmap & ideas

  • Extended case study and technical documentation (shared under NDA where applicable).
  • Generalization to other product categories and design domains.
  • Closer integration with CAD and simulation pipelines.
Back to projects