PeopleofSuperintelligence

Researcher

Neha Gupta

Neha Gupta is a machine-learning researcher whose papers cover faster language-model inference and reasoning models. Degrees from IIT Delhi and Stanford came before a computer science PhD at Stanford, completed in 2022. Faster Cascades via Speculative Decoding, co-written with Harikrishna Narasimhan, Aditya Krishna Menon and Sanjiv Kumar, earned an honorable mention in the ICLR 2025 outstanding paper awards. The later papers include Magistral, which introduces Mistral's first reasoning model, along with Voxtral and Ministral 3. The most recent listed paper, Ministral 3, was submitted to arXiv in January 2026.

Works
6
Citations
3
h-index
1

Publication numbers from OpenAlex, as of 15 October 2026.

Timeline

  1. 2022

    A Stanford computer science PhD

    After bachelor's and master's degrees in computer science from IIT Delhi and Stanford, Neha Gupta completes a computer science PhD at Stanford in 2022.

  2. June 2022

    A bias-variance view of ensembles

    Ensembling over Classifiers: a Bias-Variance Perspective is submitted to arXiv in June 2022 by Neha Gupta and three co-authors.

  3. May 2024

    Faster cascades via speculative decoding

    Faster Cascades via Speculative Decoding, co-written with Harikrishna Narasimhan, Aditya Krishna Menon and Sanjiv Kumar, is submitted to arXiv in May 2024.

  4. 2025

    An ICLR honorable mention

    ICLR's 2025 outstanding paper awards name Faster Cascades via Speculative Decoding among its honorable mentions, with Neha Gupta among the paper's authors.

  5. June 2025

    Magistral, a reasoning model

    Magistral, which introduces Mistral's first reasoning model, is submitted to arXiv in June 2025 with Neha Gupta among its authors.

  6. July 2025

    Voxtral audio chat models

    Voxtral, which presents two multimodal audio chat models, is submitted to arXiv in July 2025 with Neha Gupta among its authors.

  7. January 2026

    Ministral 3 dense language models

    Ministral 3, a family of parameter-efficient dense language models in 3B, 8B and 14B sizes, is submitted to arXiv in January 2026 with Neha Gupta among its authors.

Sources and credits

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

  1. Neha Gupta web.stanford.edu. Checked 15 October 2026.
  2. Neha Gupta - The Mathematics Genealogy Project mathgenealogy.org. Checked 15 October 2026.
  3. [2206.10566] Ensembling over Classifiers: a Bias-Variance Perspective arxiv.org. Checked 15 October 2026.
  4. [2405.19261] Faster Cascades via Speculative Decoding arxiv.org. Checked 15 October 2026.
  5. Announcing the Outstanding Paper Awards at ICLR 2025 – ICLR Blog blog.iclr.cc. Checked 15 October 2026.
  6. [2506.10910] Magistral arxiv.org. Checked 15 October 2026.
  7. [2507.13264] Voxtral arxiv.org. Checked 15 October 2026.
  8. [2601.08584] Ministral 3 arxiv.org. Checked 15 October 2026.