From Hey Siri to Hey Subbu? Apple hires Amar Subramanya, taking a bite out of Google via Microsoft

Apple has hired Google veteran Amar Subramanya as its new VP of AI to bolster its lagging ambitions. Subramanya, who previously led AI efforts at Microsoft, brings expertise in foundation models and machine learning. This move highlights the intense talent war in Silicon Valley as Apple seeks to catch up with rivals in the AI race.
TOI correspondent from Washington: In the predatory world of American artificial intelligence, where top engineers command salaries in the millions and signing bonuses rival small fortunes, Apple Inc has taken a big bite out of Microsoft and Google. The iPhone giant on Monday announced the hiring of Amar Subramanya, a Bengaluru-educated Google veteran who briefly led AI efforts at Microsoft, as its new vice president of artificial intelligence, in an effort to catch up with rivals who have sped ahead. Subramanya steps into a role vacated by the retiring John Giannandrea, inheriting oversight of Apple's foundation models, machine learning research, and AI safety teams amid a high-stakes push to salvage its lagging AI ambitions. Subramanya's whirlwind journey—from Alphabet's Mountain View labs, where he helmed engineering for the Gemini AI assistant, to Microsoft's Redmond headquarters for a five-month co-piloting stint as corporate vice president of AI, and now to sickly Siri at Apple's Cupertino campus—epitomizes the escalating talent war gripping Silicon Valley and beyond.Also read: Apple is going back to Intel after 5-long years, what makes this important for Apple and Intel With Microsoft poaching over 20 researchers from Google's DeepMind unit earlier this year, including Subramanya in July, and now Apple counter-recruiting from its rivals, the US AI ecosystem has become a revolving door of elite minds, many of them immigrants like Subramanya. The migration isn't just about prestige; it's a battle for the intellectual capital needed to build the next generation of intelligent machines, as companies like OpenAI, Meta, and Anthropic lure defectors with million dollar windfalls, even as MAGA myopics mope about “slave labor” and “code coolies” from India and wallow in toxic gibes on X. At 46, Subramanya arrives with impeccable credentials tailored to Apple's pain points. After earning a BE in electrical, electronics, and communications engineering from Bangalore University in 2001, he completed his PhD at the University of Washington in Seattle in 2009 with a dissertation on semi-supervised learning and graphical models, innovative techniques for training AI systems with limited labeled data — now a boon for privacy-obsessed Apple, which shuns vast user-data troves in favor of synthetic and licensed datasets. Subramanya's graduate work also extended to practical applications in speech recognition, natural language processing (NLP), and human activity analysis. As a visiting researcher at Microsoft Research, he tackled multi-sensory fusion for robust speech systems and speaker verification, earning a prestigious Microsoft Research Graduate Fellowship in 2007. This early exposure to Big Tech's labs foreshadowed his career, blending theory with deployment. His scholarly output includes the influential book Graph-Based Semi-Supervised Learning (co-authored with Partha Pratim Talukdar), a staple in machine learning curricula. His academic papers span entity resolution, multilingual Natural Language Processing, cross-document coreference, and audio-visual speech enhancement, underscoring his focus on scalable, multimodal systems. These foundations propelled Subramanya's 16-year Google tenure, starting around 2009 in research roles that evolved into engineering leadership. By 2023, he was head of engineering for Gemini, Google's multimodal powerhouse boasting up to 1.2 trillion parameters. Under his watch, teams of over 100 bridged experimental prototypes to production, integrating text, image, and video processing into Search, YouTube, and Android. Subramanya's pivot to Microsoft in July 2025 – along with a score of gearheads including engineering lead Sonal Gupta – was a cultural reset. In a LinkedIn post that raised plenty of eyebrows, he praised the software giant's "refreshingly low-ego yet bursting with ambition" ethos, throwing shade at Google's high-pressure environment. As corporate VP of AI, he contributed to foundation models powering Copilot, the enterprise chatbot embedded in Office and Azure.Also read: Apple, Samsung, Xiaomi, Vivo, Oppo, OnePlus and other smartphone companies must have this app pre-installed Both Microsoft and Google are led by Indian-Americans – Satya Nadella and Sundar Picha- – but Subramanya’s stint outside Seattle lasted less than six months before Tim Cook lured him back to Silicon Valley. For Apple, reeling from AI stumbles, Subramanya is a strategic antidote. The company's Apple Intelligence suite, unveiled in October 2024, has faltered: notification summaries spawned fabricated headlines, from false reports of a suspect's suicide to premature sports victories to Netanyahu’s arrest, drawing ire and lawsuits. Siri's promised overhaul—touted for contextual awareness and app control—languished after internal tests flopped, delaying rollout to spring 2026. Giannandrea, like Subramanya a Google import who once led Machine Intelligence, is said to have fostered research but struggled with product delivery. His 2024 paper on "large-reasoning models" critiqued AI's simulated cognition, yet Apple's outputs lagged. Now, with Giannandrea advising until spring, Subramanya will report to software SVP Craig Federighi, tasked with accelerating a 1-trillion-parameter in-house model while leaning on Apple’s upcoming $1 billion licensing deal with Gemini to revive Siri – delicious irony in the Apple-Google feud spanning maps to mobile.
End of Article