Aravind Srinivas
Last reviewed
May 31, 2026
Sources
No citations yet
Review status
Needs citations
Revision
v2 · 3,506 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
May 31, 2026
Sources
No citations yet
Review status
Needs citations
Revision
v2 · 3,506 words
Add missing citations, update stale details, or suggest a clearer explanation.
Aravind Srinivas (born 1994) is an Indian-American computer scientist and entrepreneur who is the co-founder and chief executive officer of Perplexity, an artificial intelligence company that operates a conversational answer engine and the Comet web browser. He founded the company in August 2022 together with Denis Yarats, Andy Konwinski, and Johnny Ho after completing a doctorate in computer science at the University of California, Berkeley under the supervision of Pieter Abbeel.[1][2]
Before launching Perplexity, Srinivas conducted research at OpenAI, DeepMind, and Google, working on policy-gradient reinforcement learning, contrastive representation learning, and vision transformers.[3][4] He returned to OpenAI as a research scientist in 2021 and contributed to work on large language and diffusion models before leaving in 2022 to found Perplexity.[3][5]
Under his leadership, Perplexity grew from a small research-driven project into one of the most highly valued private companies in artificial intelligence, reaching a reported valuation of approximately $20 billion in September 2025.[6] Srinivas has positioned the company as a direct competitor to traditional web search, has overseen the launch of the Comet browser as an "agentic" interface to the web, and has become a prominent public commentator on the future of AI-driven information retrieval.[7][8] He has also been at the center of disputes with major news publishers over the use of their content in Perplexity's answers.[9][10]
| Born | 1994, Chennai, Tamil Nadu, India |
| Nationality | Indian |
| Education | B.Tech. and M.Tech. in Electrical Engineering, Indian Institute of Technology Madras; Ph.D. in Computer Science, University of California, Berkeley (2021) |
| Doctoral advisor | Pieter Abbeel |
| Known for | Co-founding and leading Perplexity |
| Title | Co-founder and CEO, Perplexity |
| Company | Perplexity AI, Inc. (San Francisco, California) |
Srinivas was born and raised in Chennai (formerly Madras), the capital of Tamil Nadu in southern India.[11] He attended school in Chennai and prepared for the Joint Entrance Examination, the national engineering admissions test, which he has said he narrowly missed qualifying for the computer science programme at the Indian Institute of Technology Madras by a small fraction of a mark; he instead entered the electrical engineering department.[11]
From 2012 to 2017, Srinivas pursued a dual-degree programme at IIT Madras, earning a Bachelor of Technology and a Master of Technology in electrical engineering.[1][11] During the latter part of his time at IIT Madras he began research in machine learning, mentored by Professor Balaraman Ravindran of the IIT Madras Department of Computer Science, who is widely regarded as one of India's leading reinforcement-learning researchers. Despite being enrolled in the electrical engineering programme, Srinivas pursued research across departments, working on reinforcement-learning agents for Atari video games such as Pong and Breakout and on transfer learning and hierarchical reinforcement learning.[4][24] His undergraduate work resulted in papers published or co-authored at major machine-learning venues including the International Conference on Learning Representations (ICLR), the AAAI Conference on Artificial Intelligence, and the Conference on Neural Information Processing Systems (NeurIPS).[24]
In 2017 he moved to California to begin a doctorate in computer science at the University of California, Berkeley. He worked in the Berkeley Artificial Intelligence Research (BAIR) Lab under Pieter Abbeel, a professor of electrical engineering and computer sciences known for his work in robot learning and reinforcement learning.[4] His doctoral research, ultimately collected in the 2021 dissertation Representation Learning for Perception and Control, centred on self-supervised representation learning for reinforcement learning and on the application of self-attention and contrastive methods to vision and control problems.[4][25] Among the works produced during this period was the 2020 paper "CURL: Contrastive Unsupervised Representations for Reinforcement Learning," published at the International Conference on Machine Learning (ICML), which showed that contrastive learning of pixel representations could substantially improve sample efficiency in reinforcement-learning agents that operate directly from images.[25] He also contributed to vision research on Bottleneck Transformers and HaloNet during Google internships, work that helped popularise hybrid convolution–attention architectures for image recognition.[3]
He has cited Berkeley professors Abbeel and Alexei Efros, as well as fellow researcher John Schulman, as significant influences during his graduate studies.[4] In 2020 he helped co-create the Berkeley graduate course CS 294: Deep Unsupervised Learning with Abbeel, Peter Chen, and Jonathan Ho.[4]
Srinivas completed his Ph.D. in 2021.[4]
During his Ph.D. studies, Srinivas held a series of research internships at major industrial AI laboratories that shaped his later product thinking.
He spent the summer of 2018 as a research intern at OpenAI in San Francisco, where he worked under John Schulman on policy-gradient reinforcement-learning algorithms.[3][4] In 2019 he interned for five months at DeepMind in London, contributing to large-scale self-supervised contrastive representation learning and to the development of the second-generation Contrastive Predictive Coding (CPC v2) model.[3] In 2020 and 2021 he worked at Google Research on computer vision, where he co-authored papers on Bottleneck Transformers and HaloNet that introduced self-attention into convolutional architectures for image recognition.[3]
After completing his Ph.D., Srinivas joined OpenAI as a research scientist in 2021. He worked on large language models and diffusion-based generative models during a period when OpenAI was developing systems such as the DALL-E text-to-image model.[3][5]
He left OpenAI in 2022 to start a company of his own. According to interviews and the University of California, Berkeley Sutardja Center, he had repeatedly suggested to his Berkeley lab mate Peter Chen and others that they should build a startup together; he ultimately co-founded Perplexity with Denis Yarats, Andy Konwinski, and Johnny Ho.[4][2]
Perplexity AI, Inc. was incorporated in San Francisco in August 2022.[1][2] The four co-founders brought together complementary experience: Srinivas had a research background spanning OpenAI, DeepMind, and Google; Denis Yarats had been an AI research scientist at Meta and became the company's chief technology officer; Johnny Ho, a former Quora engineer and quantitative trader at Tower Research Capital, took the role of chief strategy officer; and Google Brain–trained engineer and Databricks founding team member Andy Konwinski joined as a co-founder and investor.[1][2]
The team initially built tools that allowed users to query relational databases in natural language, but quickly pivoted to a broader idea: a general-purpose answer engine that combined web search with large language models and returned synthesised answers accompanied by inline citations to the underlying web sources.[4][12] Srinivas has described the original product hypothesis as a response to a question he had repeatedly asked himself during his Ph.D.: "What if accessing information felt like talking to a personal research assistant?"[4]
Perplexity's public-facing answer engine launched on 7 December 2022.[1] The company's stated goal, according to Srinivas, was to combine "search and large language models in a way that produces answers where every part has citations," reducing hallucinations and making AI outputs verifiable.[12] The initial product served queries through a simple text interface, returning a generated answer accompanied by a numbered list of underlying web references that users could click through to verify any specific claim. This emphasis on citation became one of the company's defining product principles and a recurring talking point in Srinivas's public communication.[12][23]
The company's early growth was fuelled by viral demonstrations on Twitter and word of mouth among researchers and engineers; by February 2023 the company reported two million unique monthly visitors.[1] Initial backers in late 2022 and early 2023 included Elad Gil, Nat Friedman, Naval Ravikant, and Jeff Bezos's Bezos Expeditions; an early Series A in 2023 priced the company at roughly $150 million pre-money.[6][13]
Under Srinivas's leadership, Perplexity expanded from a single web product into a family of consumer and developer offerings:
Srinivas has described Comet as more than a browser, telling CNBC that it is meant to function as "a true personal assistant" and, in subsequent interviews, framing it as a candidate operating system for AI agents on the web.[17][7]
Perplexity has raised capital at a pace that is unusual even by the standards of recent AI startups. Major reported rounds include:
By early 2026, public filings and trade reporting placed Perplexity's enterprise valuation in the low-$20 billion range.[1]
In January 2025, as U.S. policymakers pressed for a forced sale of TikTok's U.S. operations, Perplexity submitted a proposal to merge with TikTok U.S.[19] In August 2025 the company made an unsolicited $34.5 billion all-cash offer to acquire Google's Chrome browser, which the U.S. Department of Justice had proposed as a possible antitrust remedy in its case against Alphabet. Perplexity's offer included commitments to keep Google as the default search engine in Chrome for 100 months and to invest $3 billion in Chromium over 24 months.[19] Neither bid was accepted, but both attracted substantial media attention to Srinivas and his company.
Perplexity has used commercial partnerships to broaden the distribution of its products. In July 2025 the Indian telecommunications operator Bharti Airtel announced a partnership under which it would offer a twelve-month Perplexity Pro subscription, normally priced at roughly Rs 17,000 (approximately $200), free of charge to its more than 360 million customers in India. The agreement was described by both companies as Perplexity's first partnership with an Indian telecommunications operator and was widely reported as a significant expansion of the company's footprint in one of its largest user markets.[28]
In December 2025 the Portuguese footballer Cristiano Ronaldo announced that he had invested in Perplexity and become a brand ambassador for the company. The arrangement included a co-branded landing page on the Perplexity website and was reported by Bloomberg and other outlets as one of Ronaldo's most prominent technology investments.[29]
Perplexity has been the subject of repeated complaints and lawsuits from news publishers alleging that the company uses copyrighted articles without permission.
In response, Srinivas and his colleagues launched the Perplexity Publishers' Program on 30 July 2024, which committed the company to share advertising and subscription revenue with participating publishers. Initial partners included Time, Fortune, Der Spiegel, The Texas Tribune, and Entrepreneur. The programme was expanded in December 2024 to include publishers in the United Kingdom, Japan, Spain, and Latin America.[21][22] Srinivas described the design of the programme as intended to "ensure we have a scalable and sustainable way to align incentives for all parties."[21] A spokesperson for Perplexity later told TechCrunch, in response to the New York Times suit, that "Publishers have been suing new tech companies for a hundred years, starting with radio, TV, the internet, social media, and now AI."[10]
Srinivas is among the more visible chief executives in the current generation of AI companies. He maintains an active presence on X (formerly Twitter), where he frequently comments on product launches and on the broader competitive landscape in AI search, and he has appeared in long-form interviews on a number of podcasts and broadcast programmes.
In June 2024 he was the guest on episode 434 of the Lex Fridman Podcast, a three-hour-eleven-minute conversation in which he discussed the history and future of AI, the architecture of Perplexity's answer engine, and his philosophy of company building. In that conversation he traced his intellectual lineage from his work at Berkeley to the design choices behind Perplexity, emphasising the role of inline citations as a mechanism for reducing hallucinations and improving accountability in AI outputs.[12] He has also appeared on CNBC television, including a widely circulated October 2025 interview describing Comet as a "true personal assistant," and on the All-In Podcast, where in March 2026 he drew controversy for characterising AI-driven layoffs as part of a "glorious future" that would enable new entrepreneurship.[1][17]
In September 2024, Time magazine named Srinivas to its inaugural Time 100 AI list of the most influential people in artificial intelligence. The accompanying profile noted that he was thirty years old, that Perplexity was handling roughly 230 million queries per month at the time, and that he viewed Google as structurally disinclined to cannibalise its advertising business with AI-native search, an opening he intended Perplexity to exploit.[23] He has returned repeatedly in interviews and on social media to the argument that incumbent search providers face a structural conflict between their existing advertising revenue and the lower-margin economics of AI-native answer experiences.[23][17]
Srinivas has consistently framed Perplexity's mission as building a "trustworthy" answer engine that supplies citations for every claim it makes, contrasting that approach with general-purpose chatbots that present synthesised text without sourcing.[12] He has argued in interviews and on social media that the future of consumer search is "agentic," meaning that browsers and assistants should not only retrieve information but take actions on a user's behalf, and that the Comet browser is the company's principal expression of that thesis. In discussions of Comet he has described the browser as a step toward "the OS for AI," referring to the user-facing operating environment through which people will increasingly interact with autonomous agents.[7][17]
In addition to his role at Perplexity, Srinivas has been an active angel investor since 2023, backing early-stage startups in AI and adjacent fields.[11] He has delivered guest lectures at his alma maters, including a leadership lecture at IIT Madras in early 2024, and has spoken at university and industry events on entrepreneurship in artificial intelligence.[24]
Srinivas has also engaged in extended public exchanges with executives at other AI laboratories on X, including Sam Altman of OpenAI and Elon Musk of xAI, debating product strategy, model evaluation, and the appropriate role of advertising in AI-driven search. Coverage of these exchanges in trade publications has emphasised the unusual visibility of an early-stage chief executive who routinely engages with the leaders of the largest model providers on a public platform.[17][23]
He has spoken in interviews about lessons he draws from running a fast-growing AI company in a competitive landscape dominated by much larger incumbents. Recurring themes include the value of focusing on a single, narrowly defined user problem (in his telling, "answers, with citations"), the importance of shipping products quickly and iterating with users, and the role of a research culture in maintaining technical defensibility as larger model providers commoditise individual capabilities.[4][12]
Srinivas has shared little detail about his personal life in public, and few biographical particulars beyond his birthplace, education, and career trajectory have been independently verified. He has spoken in interviews about growing up in Chennai and about the influence of his time at IIT Madras and Berkeley on his approach to building Perplexity.[4][11] He is based in San Francisco, where Perplexity is headquartered.[1]