AI in Manufacturing

Why AI in AM?

Modern additive manufacturing generates massive amounts of data, yet most decisions—from parameter selection to defect diagnosis—still rely heavily on human expertise. Our research aims to bridge this gap by developing AI systems that continuously learn from multimodal sensor data, understand the underlying manufacturing physics, and autonomously optimize printing processes. The result is a new generation of intelligent manufacturing systems capable of producing higher-quality parts with greater efficiency, reliability, and scalability.

Intelligent Monitoring

We develop computer vision and multimodal sensing algorithms that continuously monitor the printing process, detect defects as they emerge, and provide real-time insights into build quality. By integrating optical, thermal, acoustic, and photodiode data, our models enable early anomaly detection and process understanding.

Representative Publications

Self-Correcting Additive Manufacturing with Large Language Models

AI Saves 3D Prints (CMU News, 2026)
LLM-based real-time process correction

AMGPT: Large Language Models for Additive Manufacturing

Additive Manufacturing, 2025

Learning-based Defect Detection for LPBF

Melt pool monitoring
Surface defect classification
Vision-based inspection

Photodiode-Based Intelligent Monitoring

Process monitoring using optical sensing
In-situ quality prediction

Foundation Models

We develop computer vision and multimodal sensing algorithms that continuously monitor the printing process, detect defects as they emerge, and provide real-time insights into build quality. By integrating optical, thermal, acoustic, and photodiode data, our models enable early anomaly detection and process understanding.

Printer Foundation Models

Cross-machine representation learning
Manufacturing foundation models

Pretraining for Manufacturing Intelligence

Large-scale manufacturing representation learning

Transfer Learning Across AM Systems

Domain adaptation for new materials and machines

Large Language Models for Manufacturing

We develop manufacturing-specific large language models that assist engineers in process planning, parameter optimization, failure diagnosis, and knowledge retrieval. These AI copilots combine manufacturing expertise with reasoning capabilities to accelerate design, troubleshooting, and decision-making.