Akhilesh Kumar, principal research and development engineer at Ansys, has received the quarterly Silicon Integration Initiative Pinnacle Award, which recognizes volunteers for their contributions to Si2’s success as a research and development joint venture for semiconductor design automation software.
Kumar was honored for his contributions to advancing a standard application programming interface for accessing processed data from EDA tools and databases. This interface, known as the Secure ProcEssEd Data (SPEED) API, will help enable secure and uniform access to data needed to develop, train, and implement artificial intelligence and machine learning engines that accelerate chip design processes.
“EDA tools typically offer access to their design data via our OpenAccess database and API,” said Leigh Anne Clevenger, Si2 vice president of technology. “However, until now, there has not been a standard API for the processed data these tools output. As a result, users often need to create various extraction and translation scripts to convert the data into a suitable format for AI/ML training. This process is often time-consuming and error-prone, as any changes to the output file parameters or formats can break translation scripts and disrupt access to legacy training data.”
Si2’s AI/ML Special Interest Group conducted a survey identifying key obstacles to using AI/ML for advanced electronic design. The results indicated a need for a standard API for processed data and that the limited availability of training data for AI/ML models hindered significant progress in this field.
John Ellis, Si2 president and CEO, stated, “The electronics industry is facing a number of challenges, including a rapid increase in design complexity and a growing shortage of skilled talent for next-generation system design. To overcome these issues, we need to optimize current design processes and create new AI-powered tools to help navigate the complex tradeoffs necessary for optimizing system-level designs. These tools are essential to helping us understand and address the increasing complexity of designing advanced electronic systems. SPEED API is a key missing building block and its adoption will help accelerate the development of AI/ML-enhanced design flows.”
Ellis also praised Kumar’s contributions to the SPEED API project, stating, “Akhilesh has been a great catalyst for driving the SPEED API effort forward. He brings a wealth of industry experience and encourages new ideas and strategies, fostering a sense of ownership among group members. His award is well-deserved, and I’m excited to see the impact of the SPEED API under his leadership.”
Kumar’s R&D focus at Ansys is on EDA reliability solutions, including electrothermal, thermal, electrostatic discharge, and substrate noise analysis applications. His research interests are ML and algorithmic techniques for EDA solutions. He has published papers in top conferences and journals within the broad area of EDA and worked on several AI/ML projects involving multiple EDA tools with complex data interfaces and models.
Kumar previously served as a design engineer at STMicroelectronics, working on behavioral, timing, and power models for memories and memory controllers. He received his Ph.D. and M.S. in electrical and computer engineering from the University of Waterloo and a Bachelor of Technology in electrical engineering from the Indian Institute of Technology, Roorkee.