July 18, 2024. Repost from Bloomberg by Jane Lanhee Lee and Vlad Savov -- Rain AI has recruited an Apple Inc. chip executive to lead hardware engineering, the second high-profile hire this month for a startup trying to design a new type of semiconductor for artificial intelligence.
Jean-Didier Allegrucci, who helped oversee the development of processors that replaced Intel Corp. in iPhones and Macs, will steer Rain AI’s efforts to create more energy-efficient chips. He’ll work with lead architect Amin Firoozshahian, who moved over to the startup after five years at Meta Platforms Inc.
Backed by OpenAI co-founder Sam Altman and Y Combinator, San Francisco-based Rain AI is one of a growing cohort of startups looking to develop new hardware for AI platforms. It’s exploring a technique known as in-memory compute, modeled on the way the human brain works, which promises to save on power consumption by processing data where it’s stored.
Rain AI joins Intel, Taiwan Semiconductor Manufacturing Co. and Samsung Electronics Co. in researching in-memory computing, which still faces questions around economic viability and environmental impact. But in the long run, more efficient chips are essential to adding AI to a wider range of devices.
“Our novel compute-in-memory technology will help unlock the true potential of today’s generative AI models, and get us one step closer to running the fastest, cheapest, and most advanced AI models anywhere,” Rain AI Chief Executive Officer William Passo said in a statement.
Related post
Leading Chip Architect Joins Rain AI to Accelerate Vision to Enable Advanced AI Everywhere
June 4, 2024, AccessWire / Rain AI, an AI hardware startup and compute-in-memory pioneer whose mission is to enable advanced AI everywhere, today announced that former Meta architecture leader Amin Firoozshahian has joined the team as Lead Architect. In his new role, Firoozshahian will lead the design and development of new technologies that will significantly reduce the energy required to run AI inference and model training, allowing a growing list of customers and partners to deploy and productize AI.
"We could not be more excited to have an architect of Amin's caliber leading the next iteration of our architecture," said Rain AI CEO William Passo. "Our novel compute-in-memory technology will help unlock the true potential of today's generative AI models, and get us one step closer to running the most advanced models anywhere."
Firoozshahian joins the company from Meta, where he was a leading architect on the AI Systems Hardware/Software Co-Design and Infra Silicon teams working on the company's MTIA family of accelerators. In 2023, he was the lead author of Meta's 2023 paper on MTIAv1. Prior to joining Meta in 2019, he worked in Intel's Product Architecture Group, integrating novel technologies into the Xeon family of products and worked on AI accelerator performance evaluation as part of the company's Artificial Intelligence Product Group (AIPG). He joined Intel in 2014 when that company acquired Hicamp Systems Inc., a startup that developed novel memory system architectures. He received his Ph.D. in Electrical Engineering from Stanford University in 2009.
About Rain AI
Rain AI's mission is to enable advanced AI everywhere. Rain AI creates accelerator solutions for AI inference and training using novel compute-in-memory (CIM) technology, RISC-V processing cores, and specialized machine learning model optimization tools. By designing chips to run the world's most powerful AI models, Rain AI can dramatically reduce the energy consumption of AI compute. Rain AI is backed by investors including Sam Altman, Dan Gross, Grep VC, FoundersX Ventures and Y Combinator.