OpenAI’s latest GPT-4o model has wowed both the AI community and the general public alike — the model has been seen to be a significant advancement over OpenAI’s last state-of-the-art GPT-4 model. But even as indepdendent researchers are putting GPT-4o through the paces on AI benchmarks, the model has been revealed to have an Indian connection.
OpenAI CEO Sam Altman has credited Pune boy Prafulla Dhariwal for the creation of GPT-4o. “GPT-4o would not have happened without the vision, talent, conviction, and determination of Prafulla Dhariwal over a long period of time,” he posted on X. “That (along with the work of many others) led to what I hope will turn out to be a revolution in how we use computers” he added.
Prafulla Dhariwal had also posted about the creation of GPT-4o. “GPT-4o (o for “omni”) is the first model to come out of the omni team, OpenAI’s first natively fully multimodal model. This launch was a huge org-wide effort, but I’d like to give a shout out to a few of my awesome team members who made this magical model even possible!” he had said.
28-year old Prafulla Dhaliwal has been working with OpenAI since 2017, but he has his roots in India. He had been a child prodigy, having won a gold medal at the International Astronomy Olympiad in 2009 while in school in India. He had begun attending the Dr Kalmadi Shyamrao High School in Pune in 2011. While still in school, he had been training Pune’s students for the Indian National Mathematical Olympiad, and had taught algebra, functional equations and inequalities, and had designed and graded problem sets in Number Theory and Geometry. In 2013, like millions of Indian students, he’d written the IIT-JEE exam, and had secured a rank of 165.
But instead of joining one of the IITs, he had gone ahead and joined MIT in the US. At MIT, he had completed a Bachelor’s degree in Computer Science, and had also interned at some of the top companies including Pinterest and D.E. Shaw. At MIT, Dhariwal had worked at the Computer Vision Group which used Deep Learning to understand images, and at the Centre for Brain, Minds and Machines, where he worked on CNN and RNN based models for learning invariant representations for speech classification.
In 2016, Prafulla Dhariwal had joined OpenAI as a Research Intern, and had eventually joined the company as a Research Scientist in 2017. Over the years, he seems to have had a big role to play in some of the company’s biggest releases — on his X profile, he describes himself as the co-creator of GPT-3, DALL-E 2 and other OpenAI products. And now Sam Altman has credited him as being pivotal behind the creation of OpenAI’s latest model, GPT-4o.
Prafulla Dhariwal isn’t the only Indian researcher who has had a big part to play in the current AI revolution. Two Indian researchers at Google, Ashish Vaswani and Niki Parmar, had been the co-authors of the seminal “Attention is all you need” paper which introduced the concept of the transformer, which underpins many of the recent advancements in AI. Anand Srnivas, meanwhile, is the founder of the hugely successful AI company Perplexity AI, which many have dubbed to be a Google competitor. And while India is still getting up to speed with homegrown AI companies including Krutrim and Sarvam AI, Indian-origin researchers and founders are already at the forefront of the AI revolution.