PeopleofSuperintelligence

Researcher

Abhishek Sinha

Abhishek Sinha is a machine learning researcher whose work spans neural-network training, visual search, few-shot generation and multimodal systems. By 2019, they worked on a clothing visual-search product at Adobe India; at Stanford, they earned a computer science master's in 2021 while researching generative models with Stefano Ermon. Their 2021 papers explored negative data augmentation and diffusion-based few-shot generation; a paper on decision-focused distribution comparison won ICLR's 2022 Outstanding Paper recognition. At Waymo, they improved perception-model data efficiency and contributed to multimodal post-training for Gemini 1.5 and Gemini 2.5. Since April 2026, they are a Staff Software Engineer at Google DeepMind, working on Gemini image generation and Nano Banana Pro.

Badges
Links
Google Scholar and LinkedIn

Timeline

  1. 2017

    Learning weight evolution

    With Mausoom Sarkar, Aahitagni Mukherjee and Balaji Krishnamurthy, they coauthor “Introspection: Accelerating Neural Network Training By Learning Weight Evolution” in 2017. The method learns a weight-evolution pattern from MNIST classification and uses it to speed training on CIFAR-10 and ImageNet; ICLR accepts the paper in 2017.

  2. By 2019

    Visual search at Adobe

    At Adobe India, they work on a deep learning-based visual search product for clothing recommendation. In 2019, they coauthor “Powering Robust Fashion Retrieval With Information Rich Feature Embeddings” with Ayush Chopra, Hiresh Gupta, Mausoom Sarkar, Kumar Ayush and Balaji Krishnamurthy; the paper wins a Best Paper Award at a CVPR workshop.

  3. 2021

    A Stanford master's

    They earn a master's degree in computer science from Stanford University in 2021 and work as a research assistant under Professor Stefano Ermon on generative models and self-supervised learning.

  4. 2021

    Negative Data Augmentation

    In 2021, they coauthor “Negative Data Augmentation” with Stefano Ermon as first author. The paper uses intentionally out-of-distribution samples in GAN training and applies negative augmentation to contrastive learning for image and video representations.

  5. 2021

    Few-shot diffusion generation

    In 2021, they coauthor D2C with Jiaming Song, Chenlin Meng and Stefano Ermon as first author. The diffusion-decoding model pairs a learned diffusion prior with contrastive self-supervision and adapts to new generation tasks from as few as 100 labeled examples.

  6. 2022

    An ICLR Outstanding Paper

    With Shengjia Zhao, Yutong He, Aidan Perreault, Jiaming Song and Stefano Ermon, they coauthor “Comparing Distributions by Measuring Differences that Affect Decision Making.” The method compares distributions through the optimal loss for a decision task, with applications to two-sample tests, climate-change studies, sample-quality evaluation and feature selection. ICLR names the paper an Outstanding Paper in 2022.

  7. 2024–2025

    Waymo perception and Gemini reports

    Before Google, they work on Waymo's Perception Team, where they improve data efficiency for perception models. Through multimodal post-training, they contribute to Gemini 1.5 in 2024 and Gemini 2.5 in 2025; they appear among the authors of both reports.

  8. Since April 2026

    Staff engineer in image generation

    They hold a Senior Software Engineer role at Google DeepMind from January 2025 to April 2026 and become a Staff Software Engineer there in April 2026. They work on image generation from Gemini and contribute to its first native image-generation launch. For Nano Banana Pro, they are a post-training co-lead and a pre-training core contributor.

Career

  1. NowGoogle DeepMindSenior Software Engineer
  2. Stanford UniversityResearch Assistant

Education

  1. 2021Stanford UniversityMaster's degree in computer science

From the biography.

  1. 16% of AI career moves in the 2020s went from a university to a companyIncludes Abhishek Sinha. 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.
  2. Google had the most people shaping AI every year from 2014 to 2025Includes Google. Our estimates of which organizations employed the most people shaping AI each year from 1955 to 2025, tracing the shift from university labs to companies.
  3. Google has the highest combined h-index among AI organizationsIncludes Google. We estimate that Google holds the highest combined h-index at 8,016, while universities account for 73% of the deduplicated total across AI organizations.
  4. 20% of people at companies who shape AI also hold a university role, led by Google and Stanford UniversityIncludes Google. We estimate that 539 of the 6,766 people with a known current organization (8%) hold both a university role and a company role, including 20% of people at companies and 13% of people at universities.

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 Abhishek Sinha yet. Send us a free-licensed photo.

  1. Abhishek Sinha a7b23.github.io. Checked 22 October 2026.
  2. [1704.04959] Introspection: Accelerating Neural Network Training By Learning Weight Evolution arxiv.org. Checked 22 October 2026.
  3. Powering Robust Fashion Retrieval With Information Rich Feature Embeddings openaccess.thecvf.com. Checked 22 October 2026.
  4. [2102.05113] Negative Data Augmentation arxiv.org. Checked 22 October 2026.
  5. [2106.06819] D2C: Diffusion-Denoising Models for Few-shot Conditional Generation arxiv.org. Checked 22 October 2026.
  6. ICLR 2022 Awards iclr.cc. Checked 22 October 2026.
  7. [2403.05530] Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context arxiv.org. Checked 22 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 22 October 2026.
  9. Find Abhishek Sinha's email & phone number - Staff Software Engineer at Google DeepMind, Stanford | SignalHire signalhire.com. Checked 22 October 2026.