Bio

Niki Parmar is an Indian computer scientist, born in Pune, India. She is one of the eight co-authors of the June 2017 Google Brain paper "Attention Is All You Need", which introduced the transformer architecture, and a co-founder of Adept AI (2021) and Essential AI (2023). As of May 2026, she is a member of technical staff at Anthropic, the role she joined in December 2024 after departing Essential AI in September of the same year, and was a contributor to Claude 3.7.

At a glance

• Education: Bachelor of Engineering in information technology, Pune Institute of Computer Technology (2013); Master of Science in computer science, University of Southern California (2015).
• Current role: Member of Technical Staff at Anthropic since December 2024.
• Key contributions: Co-author of "Attention Is All You Need" (June 2017); contributor to "Image Transformer" and "Stand-Alone Self-Attention in Vision Models"; co-founder and Chief Technology Officer of Adept AI (2021); co-founder of Essential AI (2023); contributor to Claude 3.7 at Anthropic.
• X / Twitter: @nikiparmar09
• LinkedIn: nikiparmar

Origins

Parmar grew up in Pune, India, and pursued an engineering education at the Pune Institute of Computer Technology, where she completed a Bachelor of Engineering in information technology. Her interest in machine learning developed during her undergraduate years through Andrew Ng and Peter Norvig's introductory online courses on machine learning and artificial intelligence, which she has cited in interviews as a formative influence.

She moved to the United States for graduate studies, completing a Master of Science in computer science at the University of Southern California in 2015. During her master's program she worked in the computational social science research group led by Morteza Dehghani, applying machine-learning methods to questions in computational social science.

Career

Parmar joined Google Research in 2015 as a software engineer and transitioned to Google Brain in 2017 as a research software engineer. The defining research artifact from her Google Brain period is "Attention Is All You Need", submitted to arXiv in June 2017 and presented at NeurIPS 2017. She is one of eight authors alongside Ashish Vaswani, Noam Shazeer, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Łukasz Kaiser, and Illia Polosukhin. The paper introduced the transformer architecture, which underlies essentially every contemporary frontier large language model. Parmar is credited with contributions to model variants and the initial codebase for the paper's experiments.

Her subsequent Google Brain research extended self-attention to additional modalities and applications. The 2018 paper "Image Transformer" with Ashish Vaswani and others applied self-attention to image generation, and the 2019 paper "Stand-Alone Self-Attention in Vision Models" with Prajit Ramachandran and others investigated self-attention as the principal building block for vision architectures, work that anticipated the subsequent emergence of vision transformers.

Parmar left Google in 2021 to co-found Adept AI with Ashish Vaswani and David Luan, the former vice president of engineering at OpenAI and a former Google director. Adept positioned itself as a research-and-product company building action-taking models for enterprise software automation. Parmar served as Chief Technology Officer. She and Vaswani departed Adept in 2022 to pursue a new research-led foundation-model thesis. The senior leadership of Adept subsequently transitioned to Amazon in 2024 in a partnership-and-licensing arrangement.

In 2023, Parmar co-founded Essential AI with Ashish Vaswani, headquartered in San Francisco. The company emerged from stealth in December 2023 with a $56.5 million Series A led by March Capital with strategic investors Google, Nvidia, AMD, KB Investment, Franklin Venture Partners, and Thrive Capital. Earlier seed capital of $8.3 million from Thrive Capital had funded the 2023 stealth period.

Parmar departed Essential AI in September 2024 and joined Anthropic as a member of technical staff in December 2024, working on the Claude model line. She publicly disclosed the move in February 2025 alongside the release of Claude 3.7 Sonnet, citing contributions to its development. Vaswani has continued as Chief Executive Officer of Essential AI through the company's $175 million Series B at a $1 billion valuation in early 2026.

Affiliations

• University of Southern California: Master's student in computer science, 2013 to 2015.
• Google: Software engineer (2015 to 2017) and Google Brain research software engineer (2017 to 2021).
• Adept AI: Co-founder and Chief Technology Officer, 2021 to 2022.
• Essential AI: Co-founder, 2023 to September 2024.
• Anthropic: Member of Technical Staff, December 2024 to present.

Notable contributions

• Attention Is All You Need (June 2017). Co-author of the Google Brain transformer paper. Parmar is credited with contributions to model variants and the initial codebase for the paper's experiments.
• Image Transformer (February 2018). Co-author of the paper applying self-attention to image generation, an early extension of the transformer architecture beyond text.
• Stand-Alone Self-Attention in Vision Models (June 2019). Co-author of the paper investigating self-attention as the principal building block for vision architectures, work that anticipated the subsequent vision-transformer literature.
• Adept AI co-founding (2021). Co-founder and Chief Technology Officer of the enterprise-action-model company.
• Essential AI co-founding (2023). Co-founder of the San Francisco foundation-model company.
• Anthropic Claude 3.7 contribution (2025). Member of Technical Staff at Anthropic since December 2024, contributing to the Claude 3.7 release.

Position in the field

Parmar's career trajectory is structurally distinctive among the eight transformer-paper co-authors. The "transformer eight" pattern of post-Google research-led founder activity has produced senior research positions and founders across Character.AI, Cohere, Sakana AI, Inceptive, Adept AI, Essential AI, and Anthropic. Parmar is the only co-author of the original transformer paper to have founded two companies and subsequently joined a frontier-tier US lab as an individual researcher, with the December 2024 Anthropic transition placing her in a senior technical role at one of the three principal frontier labs.

The Google-to-founder-to-Anthropic arc reflects a broader pattern visible in the "transformer eight" cohort, in which research-led founder activity produces senior credibility that the frontier labs subsequently absorb through individual hiring or company acquisition. Industry coverage has frequently characterized Parmar as one of the most senior women in AI research, particularly within the post-Google founder cohort, with public profiles in TIME and other outlets reflecting that positioning.

Parmar's continued involvement with Essential AI as co-founder, alongside the in-progress Anthropic technical-staff role, places her in a research-and-product position within the Claude program at Anthropic. The post-2024 trajectory will be shaped by the continued evolution of the Claude model line and the broader frontier-versus-insurgent dynamics in foundation-model research.

Outlook

Open questions and watchable signals over the next 6 to 18 months:

• Continued Anthropic research output. The contribution to subsequent Claude releases beyond 3.7, particularly any research artifacts crediting Parmar.
• Senior researcher trajectory at Anthropic. Whether the member-of-technical-staff role expands into formal senior research leadership, given the depth of Parmar's prior founding and chief-technology-officer experience.
• Public commentary and conference appearances. Whether Parmar increases her public-facing visibility through technical conference appearances, interviews, and public commentary on the transformer-and-frontier-AI trajectory.
• Continued involvement with Essential AI. Any subsequent disclosed advisory or board role at Essential AI alongside the Anthropic technical-staff position.

Sources

• Niki Parmar LinkedIn profile. Current role at Anthropic and career history reference.
• Attention Is All You Need. The June 2017 arXiv paper introducing the transformer architecture.
• Image Transformer. The February 2018 paper applying self-attention to image generation.
• Stand-Alone Self-Attention in Vision Models. The June 2019 paper investigating self-attention as a vision-architecture building block.
• Niki Parmar joins Anthropic. Parmar's February 2025 X disclosure of joining Anthropic in December 2024 and contributing to Claude 3.7.
• Transformer Co-Author Niki Parmar Joins Anthropic After Founding Two AI Startups. Analytics India Magazine coverage of the Anthropic transition.
• An interview with Niki Parmar, Senior Research Scientist at Google Brain. Earlier interview on Google Brain research and the path into machine learning.
• Essential AI emerges from stealth. VentureBeat coverage of the December 2023 Essential AI Series A and stealth exit.
• Niki Parmar Google Scholar profile. Publication record and citation count.