Auburn's Electronics Packaging Research Institute leading AI revolution in semiconductor packaging

Pradeep Lall, the MacFarlane Distinguished Professor and Alumni Professor in the Department of Mechanical Engineering and director of Auburn University’s Electronics Packaging Research Institute (EPRI), is using Artificial Intelligence (AI) to take the guess-work out of the life expectancy of electronics. And to do it quickly.

Historically, engineers attempted to estimate how long a piece of hardware would last by assuming it started out in perfect condition and followed a predictable path of wear and tear. But due to the size of the simulations, the numbers for even a single scenario could take days or even weeks to crunch.

That's not ideal — not in terms of efficiency, not in terms of reality.

"We have to contend with the messy reality of physical stressors — drops and vibrations and overheating — in a way we can't always track. These conditions are often too unpredictable to capture fully in a controlled setting," said Lall, whose expertise in the field was shaped by his time as lead engineer and later as distinguished member of technical staff with Motorola.

But the core issue? Computational lag.  

"A traditional Finite Element Analysis (FEA) would take several days to make and a couple of days to run but by then, the physical state of the device may have already changed," Lall said. "It's essentially analyzing a snapshot of the past."

Introducing specialized AI-driven defect identification, manufacturing process control, and operational reliability assessment into the equation, Lall says, provides real-time feedback.

Imagine a computer chip with its own Fitbit.

"This is a massive leap forward from traditional methods because the AI doesn't need days to run a simulation — it provides answers as things are happening," Lall said. "These models can instantly identify specific failure points — tiny cracks, loose connections — and even look at a device that was turned off during a crash or a drop and reconstruct exactly how much of its life, so to speak, was used up during that impact.

"This shift toward precise asset monitoring is crucial for modernizing national resilience," he said — and it's a shift Lall’s team within EPRI is perfectly positioned to lead. 

Credited with advancements in the field of artificial intelligence applications to electronics packaging, manufacturing and reliability, Lall has been a Principal Investigator on NASA's Integrated Vehicle Health Monitoring Program. His AI solutions for feature vector identification, fault-mode classification, sensing of impending failure and remaining useful life prediction have found wide adoption for use in extreme environments. 

Lall recently partnered with the Institute of Electrical and Electronics Engineers, of which he is a fellow, to offer a professional development course for the 2025 IEEE eLearning Library titled "AI Applications in Semiconductor Packaging."

Those applications are also a research cornerstone for the newly announced NextFlex Alabama Node, a major initiative led by Auburn to strengthen the U.S. industrial base and support the Department of War. Using packaging case studies, Lall, who serves as the director of the Alabama Node, recently expounded on the core concepts and practical applications of AI-integration in a NextFlex-hosted webinar attended by packaging engineers, process specialists, reliability analysts and technical managers.

And he had plenty of case studies.

"Modern cars have significant amount of electronics content for enabling power management, guidance, navigation and control. Advanced drive assistance systems are enabled through the use of sensors and electronics to enable real-time decision framework on the automotive platform. Much of the health of the automotive systems is enabled by on-board diagnostics systems. The use of AI for prognostics health management of systems operating in extreme environments can significantly enhance operational reliability," explained Lall

"These AI models are hungry for data to train with," Lall said. “EPRI has more than 20 years’ worth of data. That is the sort of experience that can take the industry from hindsight to insight."