Researcher

Ahmed El-Kishky

Ahmed El-Kishky is an AI researcher whose work spans mining large text collections, multilingual language systems, social-network models, and AI reasoning. He received an NSF fellowship in 2013 and pursued doctoral research in data mining at UIUC. At Facebook AI he coauthored multilingual document-alignment work; at Twitter he coauthored graph and multilingual language models. At OpenAI, he helped develop o1 and coauthored research on reinforcement learning for coding and reasoning. In October 2025, he is a research lead at OpenAI who helps the team deliver an AI system that solves all 12 problems at the ICPC world finals.

Now at
OpenAI
Works
54
Citations
1,478
h-index
16
Badges

Publication numbers from OpenAlex, as of 18 October 2026.

Timeline

  1. 2013

    An NSF fellowship

    He receives the NSF Graduate Research Fellowship in 2013.

  2. 2014

    Mining topical phrases

    In 2014, he coauthors ToPMine, an efficient method for finding topical phrases in text collections.

  3. 2015

    Doctoral study at UIUC

    By 2015, he is a second-year PhD student in UIUC's Data Mining Group, advised by Jiawei Han.

  4. 2020

    Aligning multilingual documents

    In 2020, he coauthors a multilingual document-alignment method while affiliated with Facebook AI. It improves alignment results by 7% for high-resource languages, 15% for mid-resource languages, and 22% for low-resource languages.

  5. 2022–2023

    Building Twitter's social models

    In 2022, he coauthors TwHIN, which uses knowledge-graph embeddings to improve Twitter recommendation and classification tasks. In 2023, he coauthors TwHIN-BERT, a multilingual language model trained on seven billion tweets in more than 100 languages.

  6. October 2024

    Developing OpenAI o1

    In October 2024, he is an OpenAI research scientist on o1, a reasoning model he helps develop, and discusses its chain-of-thought approach with two fellow researchers.

  7. 2025

    Training models for coding contests

    In 2025, he coauthors research on reinforcement learning for coding and reasoning, comparing o1 and an early o3 checkpoint with the specialized o1-ioi system used for the 2024 International Olympiad in Informatics.

  8. October 2025

    Research lead at OpenAI

    In October 2025, he serves as OpenAI research lead and helps the team build an AI system that solves all 12 problems at the ICPC world finals.

Career

  1. 2024OpenAIResearch Scientist, Research Lead
  2. 2022–2023Twitter
  3. 2020Meta

Education

  1. University of Illinois Urbana-ChampaignPhD

Most-cited works

  1. Beyond English-Centric Multilingual Machine Translation2020 · 480 citations
  2. Scalable topical phrase mining from text corpora2014 · Proceedings of the VLDB Endowment · 211 citationsWith Jiawei Han
  3. CCAligned: A Massive Collection of Cross-Lingual Web-Document Pairs2020 · Conference on Empirical Methods in Natural Language Processing (EMNLP) · 102 citationsWith Philipp Koehn
  4. ClusType2015 · ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD) · 84 citationsWith Jiawei Han
  5. TwHIN-BERT: A Socially-Enriched Pre-trained Language Model for Multilingual Tweet Representations at Twitter2023 · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · 78 citationsWith Jiawei Han

Sources and credits

Photo: We have no published photo of Ahmed El-Kishky yet. Send us a free-licensed photo.

  1. NSF Graduate Research Fellowship | Siebel School of Computing and Data Science | Illinois siebelschool.illinois.edu. Checked 18 October 2026.
  2. Scalable Topical Phrase Mining from Text Corpora arxiv.org. Checked 18 October 2026.
  3. Data Systems - Microsoft Research: Visitors And Interns microsoft.com. Checked 18 October 2026.
  4. Main Conference Schedule-AACL-IJCNLP aacl2020.org. Checked 18 October 2026.
  5. Massively Multilingual Document Alignment with Cross-lingual Sentence-Mover's Distance arxiv.org. Checked 18 October 2026.
  6. TwHIN: Embedding the Twitter Heterogeneous Information Network for Personalized Recommendation arxiv.org. Checked 18 October 2026.
  7. TwHIN-BERT: A Socially-Enriched Pre-trained Language Model for Multilingual Tweet Representations at Twitter arxiv.org. Checked 18 October 2026.
  8. Virtual Event: Learning to Reason with LLMs - Event | OpenAI Forum forum.openai.com. Checked 18 October 2026.
  9. Competitive Programming with Large Reasoning Models arxiv.org. Checked 18 October 2026.
  10. 😺 🎙️ PODCAST: Our interview with OpenAI! theneurondaily.com. Checked 18 October 2026.