PeopleofSuperintelligence

Researcher

Amil Merchant

Amil Merchant is a machine-learning researcher whose work applies optimization and generative AI to materials science. At Harvard, he studied applied mathematics, explored machine learning for fraud detection, and coauthored a causal-inference API. In 2021, he coauthored Learn2Hop; in 2023, he coauthored Google DeepMind’s GNoME research and in 2024 served as GNoME’s technical lead. MIT Technology Review named him to its 2024 Innovators Under 35 list; in 2024, he coauthored a parallel diffusion sampler, and in 2025 he coauthored CURIE, a scientific-reasoning benchmark for language models. As of October 2026, he is a research scientist at Google.

Now at
Google
Works
22
Citations
2,734
h-index
6
Badges

Publication numbers from OpenAlex, as of 18 October 2026.

Timeline

  1. 2018–2019

    Harvard research and causal inference

    In 2018, he is an applied-mathematics concentrator at Harvard and tries machine-learning approaches to identify medical fraud during a student data challenge. In 2019, he coauthors an API that lets researchers evaluate causal-inference models and save results for future benchmarking.

  2. 2021

    Learn2Hop for rough landscapes

    In 2021, he and three collaborators introduce Learn2Hop, a meta-learned optimizer for searching rough atomic-energy landscapes and finding low-energy structures.

  3. 2023

    GNoME predicts crystal stability

    In 2023, he and Ekin Dogus Cubuk introduce GNoME, a deep-learning system that predicts crystal stability and identifies 380,000 of its 2.2 million predictions as the most stable synthesis candidates.

  4. 2024

    GNoME technical lead

    In 2024, he is GNoME’s technical lead at Google DeepMind and is named to MIT Technology Review’s Innovators Under 35 list for his materials-discovery work.

  5. 2024

    Parallel diffusion sampling

    In 2024, he coauthors Self-Refining Diffusion Samplers with Nikil Roashan Selvam and Stefano Ermon; the method parallelizes denoising and speeds generation by up to 1.7 times on a 25-step Stable Diffusion v2 benchmark.

  6. 2025

    A benchmark for scientific reasoning

    In 2025, he coauthors CURIE, an eight-task benchmark derived from about 250 scientific papers to test language models on long-context understanding and multi-step reasoning.

  7. As of October 2026

    Research scientist at Google

    As of October 2026, he is a research scientist at Google.

A language model matched the links in this timeline to people, organizations and papers (provisional, model-judged).

Career

  1. Since 2020GoogleResearch Scientist
  2. 2024Google DeepMindTechnical Lead

Education

  1. From 2018Harvard UniversityApplied Mathematics

From the biography and OpenAlex affiliations (the years of papers that list the organization), so some years are approximate.

Most-cited works

  1. Scaling deep learning for materials discovery2023 · Nature · 1,498 citationsWith Ekin Dogus Cubuk
  2. An autonomous laboratory for the accelerated synthesis of inorganic materials2023 · Nature · 1,008 citationsWith Ekin Dogus Cubuk
  3. What Happens To BERT Embeddings During Fine-tuning?2020 · BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP (BlackboxNLP) · 148 citations
  4. Scalable Diffusion for Materials Generation2023 · 28 citationsWith Ekin Dogus Cubuk
  5. VeLO: Training Versatile Learned Optimizers by Scaling Up2022 · 15 citationsWith Luke Metz

From OpenAlex, as of 18 October 2026.

  1. 13% of AI career moves in the 2020s went from a university to a companyIncludes Amil Merchant. Our estimates of how university-to-company career transitions changed by decade, including roles added alongside an earlier job.
  2. Google has the highest combined h-index among AI organizationsIncludes Google. Our estimates of combined and median h-index at AI organizations, using current affiliations and OpenAlex records.
  3. Google alumni founded or led 79 other organizations, the most among the labs we compareIncludes Google. We trace 181 organizations founded or led by AI lab alumni, and 187 career moves between the labs.
  4. 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 from 1962 to 2025, with a year-by-year chart.

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 Amil Merchant yet. Send us a free-licensed photo.

  1. Detecting deception with machine learning seas.harvard.edu. Checked 18 October 2026.
  2. Universal Causal Evaluation Engine: An API for empirically evaluating causal inference models proceedings.mlr.press. Checked 18 October 2026.
  3. Learn2Hop: Learned Optimization on Rough Landscapes arxiv.org. Checked 18 October 2026.
  4. Millions of new materials discovered with deep learning deepmind.google. Checked 18 October 2026.
  5. Amil Merchant | Innovators Under 35 innovatorsunder35.com. Checked 18 October 2026.
  6. Self-Refining Diffusion Samplers: Enabling Parallelization via Parareal Iterations proceedings.neurips.cc. Checked 18 October 2026.
  7. CURIE: Evaluating LLMs on multitask long context scientific understanding and reasoning research.google. Checked 18 October 2026.
  8. Find Amil Merchant's email & phone number - Research Scientist at Google, Stanford signalhire.com. Checked 18 October 2026.