Skip to main content

POSTECH

SoftwarePohang, South KoreaFounded 1986· One of 364 Software companies tracked by AMPulse

Develops an AI-based data-selective machine learning (DSML) framework that predicts the mechanical strength of metal LPBF parts in seconds, accounting for internal defects, to accelerate qualification for aerospace and automotive applications.

CEO / Founder
Moo Hwan Kim
Team Size
1001-5000
Stage
Active

Technology & Products

Key Products

Data-Selective Machine Learning (DSML) framework for defect-aware yield-strength prediction

Technological Advantage

Proprietary DSML framework achieves a mean prediction error of 9.51 MPa for AlSi10Mg, reported as four times more accurate than existing methods, validated in research published in Acta Materialia. The software-layer innovation is deployable across existing LPBF systems without hardware changes, offering a protectable advantage through algorithmic IP and research publications.

Differentiation

Value Proposition

Reduces the experimental burden and time for metal LPBF part qualification by providing interpretable, defect-aware yield-strength predictions with a mean error of 9.51 MPa, compressing the iterative test-build-test cycle from months to seconds.

How They Differentiate

POSTECH's DSML framework focuses specifically on interpretable, defect-aware strength prediction for metal LPBF, unlike general simulation software (ANSYS) or broader design-for-AM platforms (nTopology). It provides human-readable equations for traceable certification decisions, whereas competitors often rely on black-box AI or broader process parameter databases (Senvol).

Market & Competition

Target Customers

Aerospace and automotive manufacturers, metal AM part producers, and research institutions requiring rapid, interpretable qualification of LPBF components.

Industry Verticals

Aerospace; Automotive; Industrial Manufacturing

Competitors

ANSYS; nTopology; Senvol

Growth & Milestones

Major Milestones

Publication of DSML framework research in Acta Materialia; Validation on AlSi10Mg alloy with reported 4x accuracy improvement over existing methods

Recent coverage of POSTECH