Oren Sultan

Oren Sultan

Research Scientist at Meta Fundamental AI Research (FAIR)·PhD Candidate in Computer Science at The Hebrew University

I work on challenging reasoning problems that require abstraction and structure — augmenting large language models with symbolic algorithms to combine the best of both worlds.

  oren.sultan (at) mail.huji.ac.il

Biography

I am an AI Research Scientist at Meta, Fundamental AI Research (FAIR), working on code reasoning, understanding and generation as part of the CodeGen team. I am also a Computer Science PhD candidate (direct track) at The Hebrew University, grateful to be advised by Prof. Dafna Shahaf. My research focuses on challenging reasoning problems that require abstraction and structure — areas where current large language models still struggle. I believe that neuro-symbolic approaches, which augment LLMs with symbolic algorithms, are a promising direction, combining the best of both worlds.

Previously, I worked as an AI Researcher at Lightricks, and as a TA for Data Science under the CIDR (Center for Interdisciplinary Data Science Research). My PhD research is generously funded by the Planning and Budgeting Committee (PBC) Fellowship of the Council for Higher Education (CHE) for outstanding PhD candidates in AI and Data Science in Israel.

I completed both my MSc and BSc in Computer Science at The Hebrew University (magna cum laude) and participated in two exchange programs with scholarships for honors students: during my BSc at The University of Melbourne, and during my MSc at The Technical University of Munich. Furthermore, I graduated from The Israeli Intelligence Corps (Unit 8200) as an officer (Captain), where I founded and led tech teams of data engineers with honors. During my reserve duty, I co-founded and led special forces of researchers and data scientists. I have also gained experience from various roles in industry, from data and software engineering to consulting.

News

Publications

LLMs versus the Halting Problem: Revisiting Program Termination Prediction
Oren Sultan, Jordi Armengol-Estape, Pascal Kesseli, Julien Vanegue, Dafna Shahaf, Yossi Adi, Peter O'Hearn
ArXiv
CWM: An Open-Weights LLM for Research on Code Generation with World Models
Meta FAIR CodeGen Team, Oren Sultan
ArXiv
A Neuro-Symbolic Approach for Reliable Proof Generation with LLMs: A Case Study in Euclidean Geometry
Oren Sultan, Eitan Stern, Dafna Shahaf
ArXiv
Visual Editing with LLM-based Tool Chaining: An Efficient Distillation Approach for Real-Time Applications
Oren Sultan, Alex Khasin, Guy Shiran, Asnat Greenstein-Messica, Dafna Shahaf
EMNLP 2024
ParallelPARC: A Scalable Pipeline for Generating Natural-Language Analogies
Oren Sultan, Yonatan Bitton, Ron Yosef, Dafna Shahaf
NAACL 2024
Life is a Circus and We are the Clowns: Automatically Finding Analogies between Situations and Processes
Oren Sultan, Dafna Shahaf
EMNLP 2022
Breakpoint Transformers for Modeling and Tracking Intermediate Beliefs
Kyle Richardson, Ronen Tamari, Oren Sultan, Reut Tsarfaty, Dafna Shahaf, Ashish Sabharwal
EMNLP 2022
From Judgement's Premises Towards Key Points
Oren Sultan, Rayen Dhahri, Yauheni Mardan, Tobias Eder, Georg Groh
arXiv preprint

Invited Talks

  • Visual Editing with LLM-based Tool Chaining 🎨
    Nvidia & NLP IL - Natural Language Processing Israel Meetup, Israel 🇮🇱 (Apr 2025).
    Slides

  • Fine-Tuning Session 2 -- Real use-case from Lightricks
    AI course, Lightricks, Israel 🇮🇱 (Sep 2024).
    Slides Video

  • Fine-Tuning Large Language Models (LLMs)
    AI course, Lightricks, Israel 🇮🇱 (Sep 2024).
    Slides Video

  • Automatically Finding Analogies between Situations and Processes
    MSc graduation ceremony, The Hebrew University, Israel 🇮🇱 (June 2023).
    Slides

  • Analogies of Situations and Processes: Detection, Explanation, Generation
    SheffieldNLP, The University of Sheffield, UK 🇬🇧 (March 2023).
    Slides

  • Using LLMs for Natural-Language Analogies Generation
    BIU-NLP, Bar Ilan University, Israel 🇮🇱 (Jan 2023).
    Slides

  • Life is a Circus and We are the Clowns: Automatically Finding Analogies between Situations and Processes
    The Hebrew University, Israel 🇮🇱 (Jan 2023).
    Slides

  • Breakpoint Transformers for Modeling and Tracking Intermediate Beliefs
    Aristo team, Allen Institute for AI (AI2), Seattle, USA 🇺🇸 (Jan 2023).
    Slides

Academic Activities

Honors & Awards

  • 2024: 3-year Fellowship for outstanding PhD candidates in AI and Data Science, The Planning & Budgeting Committee (PBC) of Council For Higher Education (CHE), Israel.

  • 2022: Erasmus+ Scholarship for honors students, Technical University of Munich, Germany.

  • 2019: B.Sc in Computer Science – Magna Cum Laude, The Hebrew University of Jerusalem, Israel.

  • 2018: Scholarship for honors students, The University of Melbourne, Australia.

  • 2013: Annual Excellence Tech Project Award, The Israeli Intelligence Corps, Commander of Unit 8200 (rank: Brigadier general).

  • 2012: Annual Excellence Team Award, The Israeli Intelligence Corps, Commander of R&D, Unit 8200 (rank: Colonel).