TCS rolls out AI solution to detect anomalies during semiconductor manufacturing

TCS rolls out AI solution to detect anomalies during semiconductor manufacturing
Photo Credit: 123RF.com
5 May, 2020

Tata Consultancy Services (TCS) has launched TCS WaferWise, a cloud-based solution to detect wafer anomalies in semiconductor manufacturing.

The Mumbai-based information technology (IT) firm’s solution leverages custom artificial intelligence (AI) models to automatically detect and classify anomalies by analysing nano-scale images generated during the semiconductor manufacturing process, the company said in a statement.

Several companies still run human-reliant manual inspections, which could make the process more prone to errors, it said. Quality inspection of wafers, which are thin slices of semiconductor substances used to make integrated circuits, is essential to detect and classify defects early and accurately, as well as improve quality, it added.

“TCS WaferWise is a transformative solution that we are deploying at the core of our customers’ manufacturing processes to improve product quality, drive up throughput and enhance their business growth,” V Rajanna, global head of technology business unit at TCS, said.

The solution runs on Google Cloud, and utilises TCS’ digital learning platform to sample different datasets across multiple manufacturing lines and enhance accuracy.

“Semiconductor companies, like all manufacturers, are looking to optimise processes, minimise costs and improve quality control. TCS is leveraging its industry expertise and Google Cloud’s infrastructure and AI capabilities to help this set of key customers modernise and streamline critical processes,” Victor Morales, managing director of partnerships at Google Cloud, said.

Recently, TCS Digitate, the software division of the company, launched an AI-powered tool to help enterprises monitor purchase transactions. Called ignio Cognitive Procurement, the solution helps companies make smarter purchase decisions and take actions faster.

Comment(s)