Latest news with #PraneetSingh

Business Standard

Trump's visa curbs vs America's need for immigrant tech brains: Decoded

Meta's latest artificial intelligence venture, the Superintelligence Lab, is being led entirely by immigrant researchers, even as US visa policies remain unpredictable for foreign professionals. The team of 11 researchers, hired from top firms like OpenAI, DeepMind, Google and Anthropic, hails from India, China, South Africa, the UK and Australia. Each holds advanced degrees from US universities and has worked on some of the most sophisticated AI systems in use today. This comes at a time when the US government has toughened its stance on immigration. The Trump administration has introduced measures like prolonged visa processing times, heightened scrutiny of social media activity, and targeted ICE enforcement across cities with large immigrant populations. Immigrant expertise behind major AI breakthroughs From developing core algorithms to building multimodal and generative AI systems, the researchers joining Meta have worked at the heart of global AI innovation. 'Meta's Superintelligence Lab, led entirely by immigrant researchers, is the practical reality we needed to see. AI innovation is truly going borderless and India continues to serve as a strong hub of ready talent,' Praneet Singh, AVP – University Partnership at upGrad's study abroad division told Business Standard. 'It's quite paradoxical, though — while our talent is ready, mobility remains constrained. Despite policy shifts like STEM OPT extensions and improved green card processing, visa constraints and bureaucratic uncertainties continue to slow down global movement,' he added. AI talent is ready—but mobility is not Neeti Sharma, CEO of TeamLease Digital, said immigrant talent has long powered US tech leadership. 'It is not just about filling job vacancies—it's about bringing in diverse skill sets, expertise and perspectives that collaboratively bring out the best for any organisation,' she told Business Standard. But she pointed out that barriers to movement are growing. 'Delayed visa processing remains the biggest hurdle: long waits, restrictive caps, and policy uncertainty disrupt both career plans and hiring strategies,' she said. She added that talent also struggles with issues like cultural integration, credential recognition gaps, and stiff competition, which can slow down or discourage Indian AI professionals—even though they remain central to innovation hubs like Silicon Valley. India's AI education still catching up While India produces a large number of engineering graduates every year, experts say the country still lacks a consistent model for preparing them for top AI roles abroad. 'On the supply side, Indian universities have made commendable strides in STEM education, but AI roles today demand more than academic depth—they require hands-on, real-time problem-solving,' said Singh. Sharma agreed. 'While our institutions are making progress, we are still far away from creating a systemic model of AI learning and upskilling our new and existing workforce,' she said. 'To truly prepare students for global AI roles, universities need stronger industry ties, updated curricula, and more emphasis on applied research. Some institutes are innovating impressively—but scaling these nationally still remains a challenge,' Sharma added. Why the US risks losing AI talent While the US still leads in AI research demand, Sharma warned that current visa and policy delays could push talent elsewhere. 'US continues to top the table on requirements for AI talent. However, their policies, especially lately, haven't kept pace with the increase in demand. In fact, longer visa timelines and uncertainty around renewals have kept many tech companies on the edge,' she said. 'Countries like Canada and the UK have moved faster to build systems that actively attract tech workers. The US still has a strong pull, but there's room to make the path more predictable and welcoming for long-term contributors, especially from countries like India,' added Mayank Kumar, co-founder and CEO of BorderPlus. Meet Meta's AI dream team Meta's new hires include some of the most cited researchers and engineers in the AI field. They have contributed to systems like GPT-4o, Gemini, and foundational models used across text, voice, image and robotics. A few notable members of the lab: Trapit Bansal (India) – PhD from University of Massachusetts Amherst, known for work in reinforcement learning and chain-of-thought reasoning. Shuchao Bi (China) – PhD from UC Berkeley, co-developed the voice mode for GPT-4o. Huiwen Chang (China) – PhD from Princeton, led image generation for GPT-4o and developed Muse architecture. Ji Lin (China) – PhD from MIT, helped scale GPT-4o and o-series models with a focus on efficiency. Joel Pobar (Australia) – former infrastructure specialist at Meta and Anthropic, contributed to systems like HHVM and PyTorch. Jack Rae (UK) – PhD from UCL, developed Google's Gopher and Chinchilla models, worked on Gemini 2.5. Hongyu Ren (China) – PhD from Stanford, worked on robustness and post-training of o-series models. Johan Schalkwyk (South Africa) – Google Fellow, early developer of voice tech in Google's Sesame project. Pei Sun (China) – developed AI perception systems at Waymo, contributed to Gemini post-training. Jiahuai Yu (China) – PhD from UIUC, focused on multimodal learning across GPT-4 and o-series. Shengjia Zhao (China) – PhD from Stanford, co-created ChatGPT and GPT-4, known for work in AI safety and data synthesis. Each of them joins Meta at a time when tech companies are racing to build next-generation AI platforms, but also when immigration pathways remain anything but certain.