avatar

Parshin Shojaee

Ph.D. Student
Virginia Tech
parshinshojaee@vt.edu


About Me

Hi! I’m a final-year PhD student in Computer Science at Virginia Tech, advised by Prof. Chandan Reddy. My research interests center on understanding how language models reason, adapt to new scenarios, and how these capabilities can be harnessed toward scientific and open-ended discovery. I mostly focus on combining techniques from natural language processing and program synthesis, developing better algorithms for language models to handle symbolic reasoning/discovery tasks, while studying their current limitations.

During my PhD, I’ve been fortunate to have my work recognized through fellowships and awards, including the Pratt Fellowship, and two papers nominated as Oral at ICML 2025 (top 1%) and ICLR 2025 (top 1.8%). My research has also been featured in Nature, Women in AI Research Podcast, CNBC, HackerNews, Forbes, Mashable, The Guardian, etc. I was fortunate to have the chance of working with brilliant researchers in the field whose mentorship and collaboration has shaped my research journey: Chandan Reddy (VT), Mehrdad Farajtabar (Apple), Samy Bengio (Apple), Iman Mirzadeh (Apple), Amir Barati Farimani (CMU), Kazem Meidani (CMU), Yuanqi Du (Cornell), Shashank Gupta (AI2), Yunyao Li (Adobe), and Khoa Doan (VinUni).

Before starting my PhD, I completed my MS in Operations Research at Virginia Tech, and obtained my BS from Sharif University of Technology.

If you’re interested in my research on open-ended reasoning and discovery with LLMs, would like to discuss relevant topics, or explore potential collaborations, please feel free to get in touch :) - I am best reached by email at parshinshojaee@vt.edu

News

Selected Publications

  1. Preprint
    Evaluating Large Language Models in Scientific Discovery
    Zhangde Song, Jieyu Lu, Yuanqi Du, Botao Yu*, Thomas M. Pruyn*, Yue Huang*, Kehan Guo*, Xiuzhe Luo*, Yuanhao Qu*, Yi Qu*, Yinkai Wang*, Haorui Wang*, Jeff Guo*, Jingru Gan*, Parshin Shojaee*, ..., Huan Sun, Mohammad Moosavi, Chenru Duan.
    arXiv preprint arXiv:2512.15567
    PDF Code Contribution

  2. NeurIPS 2025
    The Illusion of Thinking? Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity
    Parshin Shojaee, Iman Mirzadeh, Keivan Alizadeh, Max Horton, Samy Bengio, Mehrdad Farajtabar
    Advances in Neural Information Processing Systems (NeurIPS 2025).
    PDF Poster Presentation HackerNews Forbes Mashable

  3. ICML 2025
    LLM-SRBench: A New Benchmark for Scientific Equation Discovery with Large Language Models
    Parshin Shojaee, Ngoc-Hieu Nguyen, Kazem Meidani, Amir Barati Farimani, Khoda D Doan, Chandan K. Reddy
    International Conference on Machine Learning (ICML 2025).
    PDF Code Benchmark Data Oral (top 1%)

  4. ICLR 2025
    LLM-SR: Scientific Equation Discovery via Programming with Large Language Models
    Parshin Shojaee, Kazem Meidani, Shashank Gupta, Amir Barati Farimani, Chandan K. Reddy
    International Conference on Learning Representations (ICLR 2025).
    PDF Code Oral (top 1.8%)

  5. AAAI 2025
    Towards Scientific Discovery with Generative AI: Progress, Opportunities, and Challenges
    Chandan K. Reddy, Parshin Shojaee
    The Association for the Advancement of Artificial Intelligence (AAAI 2025).
    PDF

  6. ICLR 2024
    SNIP: Bridging Mathematical Symbolic and Numeric Realms with Unified Pre-training
    Kazem Meidani*, Parshin Shojaee*, Chandan K. Reddy, Amir Barati Farimani (*Equal Contribution)
    International Conference on Learning Representations (ICLR 2024).
    PDF Code Spotlight (top 5%)

  7. NeurIPS 2023
    Transformer-based Planning for Symbolic Regression
    Parshin Shojaee*, Kazem Meidani*, Amir Barati Farimani, Chandan K. Reddy (*Equal Contribution)
    Advances in Neural Information Processing Systems (NeurIPS 2023).
    PDF Video Code

  8. TMLR
    Execution-based Code Generation using Deep Reinforcement Learning
    Parshin Shojaee, Aneesh Jain, Sindhu Tipirneni, Chandan K. Reddy
    Transactions on Machine Learning Research (TMLR), 2023.
    PDF Video Code

Services

Conference Reviewer: ICLR 2026, NeurIPS 2025 (Top Reviewer), ICML 2025, AISTAT 2025, ICLR 2025, NeurIPS 2024 (Top Reviewer both Main & DB Tracks)

Journal Reviewer: Nature, Nature Machine Intelligence, TMLR

Workshop Reviewer: NeurIPS 2025 Math-AI, NeurIPS 2025 AI for Science, NeurIPS 2024 Math-AI, ICML 2024 AI for Science

A little more about me

When I’m not into AI research, I enjoy exploring all kinds of music and playing Tar and Setar instruments (Persian guitars with their own unique twist). Recently, I’m also into hiking new trails, and trying to be good at gardening :D


Powered by Jekyll and Minimal Light theme.