PeopleofSuperintelligence

Researcher

Archit Sharma

Archit Sharma is a machine learning researcher and a co-author of Direct Preference Optimization (DPO), a method for aligning language models with human preferences. After studying electrical engineering at IIT Kanpur, he worked as an AI resident at Google Brain from 2018 to 2020 and was a PhD student at Stanford from 2021 to 2024, advised by Chelsea Finn. His first-author papers on skill discovery (DADS, 2019) and autonomous reinforcement learning (VaPRL, 2021) came before DPO, which appeared in May 2023 and was named an Outstanding Paper Runner-up at NeurIPS 2023. He is a research scientist at Google DeepMind, which he joined in September 2024.

Badges

Timeline

  1. 2014–2018

    Engineering at IIT Kanpur

    He studies electrical engineering at IIT Kanpur from 2014 to 2018, with minors in machine learning and linguistic theory, and works with Piyush Rai. In 2017, he interns with Yoshua Bengio at Mila.

  2. 2018–2020

    AI resident at Google Brain

    From July 2018 to December 2020, he is an AI resident at Google Brain in Mountain View.

  3. 2019

    Skill discovery with DADS

    In July 2019, he posts Dynamics-Aware Unsupervised Discovery of Skills as first author, with Shixiang Gu, Sergey Levine, Vikash Kumar and Karol Hausman. The DADS algorithm discovers predictable behaviors and learns their dynamics at the same time.

  4. 2021–2024

    Stanford PhD under Chelsea Finn

    He is a machine learning PhD student at Stanford University from January 2021 to September 2024, advised by Chelsea Finn. His work centers on autonomous deep reinforcement learning, in which real-world robots learn to deal with unseen situations without human intervention.

  5. 2021

    Subgoal curricula and VaPRL

    In July 2021, he posts Autonomous Reinforcement Learning via Subgoal Curricula as first author, with Abhishek Gupta, Sergey Levine, Karol Hausman and Chelsea Finn. Its VaPRL method reduces the interventions required by three orders of magnitude compared to episodic RL.

  6. May 2023

    Direct Preference Optimization

    In May 2023, he coauthors Direct Preference Optimization: Your Language Model is Secretly a Reward Model with Rafael Rafailov, Eric Mitchell, Stefano Ermon, Christopher D. Manning and Chelsea Finn. DPO solves the standard RLHF problem with only a simple classification loss and eliminates the need to sample from the language model during fine-tuning.

  7. 2023

    A NeurIPS runner-up paper award

    NeurIPS 2023 names Direct Preference Optimization an Outstanding Paper Runner-up. The award recognizes the DPO paper he wrote with Rafael Rafailov, Eric Mitchell, Christopher D. Manning, Stefano Ermon and Chelsea Finn.

  8. Since 2024

    Research scientist at Google DeepMind

    He joins Google DeepMind as a research scientist in September 2024. In July 2025, he is among the authors of the Gemini 2.5 paper.

Career

  1. Since 2024Google DeepMindResearch Scientist
  2. 2018–2020Google BrainAI Resident
  3. 2017MilaIntern

Education

  1. 2024Stanford UniversityPhD
  2. 2018Indian Institute of Technology Kanpur

From the biography.

  1. Of 943 people shaping AI who worked at Google, 446 work there nowIncludes Archit Sharma. Our estimates of where the 943 people shaping AI who worked at Google and Google DeepMind work now: 446 at Google and 452 across 278 other organizations.
  2. The median h-index is 48 at universities and 14 at frontier AI labsIncludes Archit Sharma. The median h-index is 48 at universities and 14 at frontier AI labs. We compare publishing careers, recent citations, dual appointments and job moves across 2,420 university researchers and 308 frontier-lab researchers.
  3. 16% of AI career moves in the 2020s went from a university to a companyIncludes Archit Sharma. We estimate that 16% of AI career moves in the 2020s went from a university to a company; 29% of moves starting at a university went to a company, and 76% ended the university role.
  4. 23% of people shaping AI who left a university for a company later returned to a universityIncludes Archit Sharma. We estimate that 468 of the 2,011 people shaping AI who left a university for a company (23%) later returned to a university after a median of 5 years, and 163 of 829 (20%) made the reverse round trip.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Archit Sharma yet. Send us a free-licensed photo.

  1. Archit Sharma architsharma97.github.io. Checked 21 October 2026.
  2. Archit Sharma | alphaXiv alphaxiv.org. Checked 21 October 2026.
  3. Archit Sharma, Stanford: On unsupervised and autonomous reinforcement learning - Imbue imbue.com. Checked 21 October 2026.
  4. [1907.01657] Dynamics-Aware Unsupervised Discovery of Skills arxiv.org. Checked 21 October 2026.
  5. [2107.12931] Autonomous Reinforcement Learning via Subgoal Curricula arxiv.org. Checked 21 October 2026.
  6. [2305.18290] Direct Preference Optimization: Your Language Model is Secretly a Reward Model arxiv.org. Checked 21 October 2026.
  7. NeurIPS 2023 Awards neurips.cc. Checked 21 October 2026.
  8. [2507.06261] Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities arxiv.org. Checked 21 October 2026.