PeopleofSuperintelligence

Researcher

Adil Salim

Adil Salim is a researcher whose work spans language models, optimization and diffusion theory. His early career included master’s degrees in Paris, a machine-learning Ph.D. at Télécom ParisTech, a postdoctoral role at KAUST and a research fellowship at the Simons Institute. His Phi-series work includes a contribution to Phi-2 and coauthorship of the Phi-3 and Phi-4 reports; Phi-4 is a 14-billion-parameter language model. As of October 2026, he works at an early-stage startup, focusing on language-model reasoning and diffusion theory; a paper he coauthored appears as an oral at ICML 2026.

Now at
early-stage startup
Works
40
Citations
626
h-index
11
Badges

Publication numbers from OpenAlex, as of 18 October 2026.

Timeline

  1. 2015–2018

    A machine-learning doctorate

    In 2015, he earns master’s degrees in data science and probability theory at ENSAE ParisTech and Paris XI, then completes a machine-learning Ph.D. at Télécom ParisTech in 2018.

  2. 2019–2021

    Postdoctoral research at KAUST

    At KAUST, he works with Professor Peter Richtárik as a postdoctoral research fellow from 2019 to 2021.

  3. 2021

    A Simons Institute fellowship

    In Fall 2021, he joins the Geometric Methods in Optimization and Sampling program as a Simons-Berkeley research fellow.

  4. 2023

    Contributing to Phi-2

    He contributes to Microsoft Research’s Phi-2 release in December 2023, a 2.7-billion-parameter language model with reasoning and language-understanding capabilities.

  5. 2024

    The Phi-3 and Phi-4 reports

    His Phi-series work includes coauthorship of Microsoft Research’s Phi-3 report in August 2024 and its Phi-4 report in December 2024; Phi-4 is a 14-billion-parameter language model.

  6. 2026

    Dimension-free diffusion sampling

    With Khashayar Gatmiry and Sitan Chen, he coauthors “High-accuracy and dimension-free sampling with diffusions,” which appears as an oral paper at ICML 2026. It proposes a new solver with a high-accuracy guarantee whose complexity depends on dimension through the target distribution’s effective radius.

  7. As of October 2026

    Research at a startup

    He works at an early-stage startup after serving as a senior researcher at Microsoft Research. His current focus includes reasoning with language models, especially Phi, and diffusion models theory.

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

Career

  1. Nowearly-stage startup
  2. 2022–2023Microsoft ResearchSenior researcher
  3. 2022–2024Microsoft
  4. 2021Simons Institute for the Theory of ComputingSimons-Berkeley research fellow
  5. 2019–2021King Abdullah University of Science and TechnologyPostdoctoral research fellow
  6. 2016–2019Télécom Paris
  7. 2016–2019ParisTech
  8. 2016–2019Université Paris-Saclay
  9. 2016–2019Institut Polytechnique de Paris

Education

  1. 2018Télécom ParisTechPhD
  2. 2015National School of Statistics and Economic AdministrationMaster’s degree
  3. 2015Université Paris-SudMaster’s degree

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

Most-cited works

  1. Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone2024 · 164 citationsWith Ahmed Awadallah, Ronen Eldan, Shital Shah, Sébastien Bubeck and Yin Tat Lee
  2. Textbooks Are All You Need2023 · 138 citationsWith Adam Tauman Kalai, Ronen Eldan, Shital Shah, Sébastien Bubeck and Yin Tat Lee
  3. Maximum Mean Discrepancy Gradient Flow2019 · 61 citationsWith Arthur Gretton
  4. Phi-4 Technical Report2024 · 31 citationsWith Ronen Eldan, Shital Shah, Sébastien Bubeck and Yin Tat Lee
  5. A Non-Asymptotic Analysis for Stein Variational Gradient Descent2020 · 29 citationsWith Arthur Gretton

From OpenAlex, as of 18 October 2026.

  1. 13% of AI career moves in the 2020s went from a university to a companyIncludes Adil Salim. Our estimates of how university-to-company career transitions changed by decade, including roles added alongside an earlier job.
  2. The company share of AI researchers’ top papers rose from 2% to 33%Includes Microsoft. We estimate where leading AI researchers worked when their most-cited papers appeared. The company share rose from 2% in 1975–1979 to 33% in 2020–2024.
  3. Google has the highest combined h-index among AI organizationsIncludes Microsoft. Our estimates of combined and median h-index at AI organizations, using current affiliations and OpenAlex records.
  4. Google alumni founded or led 77 other organizations, the most among the labs we compareIncludes Microsoft. We trace 177 organizations founded or led by AI lab alumni, and 186 career moves between the labs.

Colleagues

People with dated positions at the same organization in the same years.

Sources and credits

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

  1. Adil Salim simons.berkeley.edu. Checked 18 October 2026.
  2. Adil Salim | Postdoctoral Research Fellow, Computer Science cemse.kaust.edu.sa. Checked 18 October 2026.
  3. Phi-2: The surprising power of small language models microsoft.com. Checked 18 October 2026.
  4. Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone microsoft.com. Checked 18 October 2026.
  5. Phi-4 Technical Report microsoft.com. Checked 18 October 2026.
  6. High-accuracy and dimension-free sampling with diffusions proceedings.mlr.press. Checked 18 October 2026.
  7. Adil Salim, Khashayar Gatmiry, Sitan Chen · High-accuracy and dimension-free sampling with diffusions · SlidesLive slideslive.com. Checked 18 October 2026.
  8. Homepage adil-salim.github.io. Checked 18 October 2026.