PeopleofSuperintelligence

Researcher

Aditya Krishna Menon

Aditya Krishna Menon is a machine learning researcher who designs and analyzes learning algorithms for problems with noisy or imbalanced labels and, more recently, for efficient inference with large language models. Born and raised in India, he earned honours in computer science at the University of Sydney in 2006 and a PhD at UC San Diego in 2013 under Charles Elkan. Between 2013 and 2018 he held research positions at NICTA and CSIRO Data61, and in 2018 he became a Fellow at the Australian National University. He coauthored a loss-correction method for noisy labels at CVPR 2017, Long-tail learning via logit adjustment at ICLR 2021, and Faster Cascades via Speculative Decoding, which received an Honorable Mention at ICLR 2025. At Google in New York, he is a research scientist.

Now at
Google
Badges
Links
Website, Google Scholar and LinkedIn

Timeline

  1. 2006

    Honours in Computer Science

    Born and raised in India, he completes an honours degree in Computer Science at the University of Sydney in 2006, under Sanjay Chawla.

  2. 2013

    A PhD at UC San Diego

    He completes a PhD in Computer Science at UC San Diego in 2013 under Charles Elkan, with the dissertation Latent feature models for dyadic prediction. His doctoral research covers dyadic prediction, the problem of predicting outcomes from the interaction of pairs of entities, with collaborative filtering and link prediction as special cases.

  3. 2013–2018

    NICTA, Data61 and ANU

    From May 2013 to June 2016 he is a researcher in the machine learning group at NICTA, and from July 2016 to December 2017 a senior research scientist at CSIRO Data61. In January 2018 he becomes a Fellow at the Australian National University, working with Lexing Xie and Bob Williamson.

  4. 2017

    Loss correction for label noise

    With Giorgio Patrini, Alessandro Rozza, Richard Nock and Lizhen Qu, he coauthors a method for training deep neural networks under class-dependent label noise, presented as an oral paper at CVPR 2017. The two loss-correction procedures amount to at most a matrix inversion and multiplication.

  5. 2021

    Long-tail learning via logit adjustment

    With Sadeep Jayasumana, Ankit Singh Rawat, Himanshu Jain, Andreas Veit and Sanjiv Kumar, he publishes this ICLR 2021 paper on classification with long-tailed label distributions. It revisits logit adjustment based on label frequencies, applied after training or enforced in the loss during training.

  6. May 2024

    Faster cascades via speculative decoding

    In May 2024, he and six coauthors post Faster Cascades via Speculative Decoding on arXiv. Their speculative cascades implement a cascade's deferral rule through speculative execution, and experiments with Gemma and T5 models show better cost-quality trade-offs than cascading and speculative decoding baselines.

  7. 2025

    An ICLR 2025 honorable mention

    ICLR names Faster Cascades via Speculative Decoding one of three Honorable Mentions among its 2025 outstanding papers, announced in April 2025. In September 2025, he and Hari Narasimhan describe the approach in a Google Research blog post on combining speculative decoding with standard cascades.

  8. As of October 2026

    Research scientist at Google

    He is a research scientist at Google in New York. His current interests include efficient inference for large language models through knowledge distillation and model cascading, along with retrieval and re-ranking.

Career

  1. NowGoogleResearch Scientist
  2. From 2018Australian National UniversityFellow
  3. 2016–2017CSIROSenior Research Scientist
  4. 2013–2016National ICT AustraliaResearcher

Education

  1. 2013University of California, San DiegoPhD
  2. 2006University of SydneyHonours in Computer Science

From the biography.

  1. 16% of AI career moves in the 2020s went from a university to a companyIncludes Aditya Krishna Menon. 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 Aditya Krishna Menon yet. Send us a free-licensed photo.

  1. Homepage of Aditya Krishna Menon akmenon.github.io. Checked 21 October 2026.
  2. Homepage of Aditya Krishna Menon (UC San Diego) cseweb.ucsd.edu. Checked 21 October 2026.
  3. Aditya Krishna Menon - The Mathematics Genealogy Project mathgenealogy.org. Checked 21 October 2026.
  4. Homepage of Aditya Krishna Menon (Australian National University) users.cecs.anu.edu.au. Checked 21 October 2026.
  5. Making Deep Neural Networks Robust to Label Noise: a Loss Correction Approach arxiv.org. Checked 21 October 2026.
  6. Long-tail learning via logit adjustment arxiv.org. Checked 21 October 2026.
  7. Faster Cascades via Speculative Decoding arxiv.org. Checked 21 October 2026.
  8. Announcing the Outstanding Paper Awards at ICLR 2025 – ICLR Blog blog.iclr.cc. Checked 21 October 2026.
  9. Speculative cascades — A hybrid approach for smarter, faster LLM inference research.google. Checked 21 October 2026.