PeopleofSuperintelligence

Researcher

Shashank Gupta

Shashank Gupta is a researcher at the Allen Institute for AI (Ai2) in Seattle, where they develop agentic language models that plan, reason and use tools. They studied computer science at the Birla Institute of Technology and Science, Pilani, and the University of Illinois at Urbana-Champaign, then worked as an applied scientist at Microsoft AI from 2018 to 2022. In 2023, they coauthored the Self-Refine method and were first author of 'Bias Runs Deep', a study of reasoning bias in persona-assigned LLMs. Their coauthored benchmarks AppWorld and SUPER won an ACL 2024 Best Resource Paper award and an EMNLP 2024 Outstanding Paper award, and in 2025 they were a core contributor to the fully open Olmo 3 language models.

Works
100
Citations
2,671
h-index
27
Badges

Publication numbers from OpenAlex, as of 22 October 2026.

Timeline

  1. 2008–2017

    Studies at BITS Pilani and Illinois

    They study for a computer science BE at the Birla Institute of Technology and Science, Pilani, from 2008 to 2012. After a research internship in the Databases and Information Systems group at the Max Planck Institute in 2014 and 2015, they study for a computer science MSc at the University of Illinois at Urbana-Champaign from 2015 to 2017.

  2. 2018–2022

    Applied scientist at Microsoft AI

    They work as an applied scientist at Microsoft AI from 2018 to 2022, leading Smart Reply research for Outlook and Teams in the Microsoft Search, Assistant and Intelligence group. Their work there covers model compression and mixture-of-experts architectures.

  3. March 2023

    Co-developing Self-Refine

    They coauthor Self-Refine, posted on arXiv in March 2023. A single language model writes an output, gives feedback on it and refines it, and across the evaluated tasks the outputs improve by about 20% absolute on average.

  4. 2023

    Bias Runs Deep at Ai2

    In November 2023, they are first author of 'Bias Runs Deep: Implicit Reasoning Biases in Persona-Assigned LLMs' with colleagues at the Allen Institute for AI. The study tests 24 reasoning datasets, 4 LLMs and 19 personas, finds that persona assignment can lower reasoning performance, and appears at ICLR 2024.

  5. 2024

    Two award-winning agent benchmarks

    AppWorld, a benchmark of 750 tasks for interactive coding agents in a simulated world of 9 apps, wins a Best Resource Paper award at ACL 2024. SUPER, a benchmark of problems on setting up and executing tasks from research repositories, wins an Outstanding Paper award at EMNLP 2024. They are among the authors of both.

  6. 2025

    LLM-SR as an ICLR oral

    They coauthor LLM-SR, a method for discovering scientific equations by programming with large language models, which is presented as an oral paper at ICLR 2025.

  7. 2025

    Core contributor to Olmo 3

    In 2025, Ai2 releases Olmo 3, a family of fully open language models at the 7B and 32B parameter scales, and posts the technical report on arXiv in December 2025. They are a core contributor to the project.

  8. As of October 2026

    Researcher at Ai2

    As a researcher at the Allen Institute for AI (Ai2) in Seattle, they develop agentic large language models: systems that plan, reason, use tools and APIs, write code and interact with complex environments.

Career

  1. Since 2024Allen Institute for AI (Ai2)Researcher
  2. 2018–2022MicrosoftApplied Scientist, Microsoft Search, Assistant and Intelligence group
  3. 2014–2015Max Planck Institute for Intelligent SystemsResearch intern
  4. 2014–2018National Institute of Technology Kurukshetra
  5. 2003–2025Birla Institute of Technology and Science, Pilani

Education

  1. 2017University of Illinois Urbana-ChampaignComputer science MSc
  2. 2012Birla Institute of Technology and Science, PilaniComputer science BE

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

Most-cited works

  1. A survey on the role of Internet of Things for adopting and promoting Agriculture 4.02021 · Journal of Network and Computer Applications · 283 citations
  2. Role of emerging technologies in future IoT-driven Healthcare 4.0 technologies: a survey, current challenges and future directions2021 · Journal of Ambient Intelligence and Humanized Computing · 238 citations
  3. Security of Cryptocurrencies in blockchain technology: State-of-art, challenges and future prospects2020 · Journal of Network and Computer Applications · 225 citations
  4. Cross-Site Scripting (XSS) attacks and defense mechanisms: classification and state-of-the-art2015 · International Journal of Systems Assurance Engineering and Management · 222 citations
  5. Role of machine learning and deep learning in securing 5G-driven industrial IoT applications2021 · Ad Hoc Networks · 142 citations

From OpenAlex, as of 21 October 2026.

  1. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes Shashank Gupta. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other university. We compare doctorate and 5+ year employment links, career timing and current workplaces for 187 major-prize and 1,111 best-paper winners.
  2. The company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes Allen Institute for AI (Ai2). We estimate where the people shaping AI worked when their most-cited papers appeared: the company share was 31% in 2020–2024 and 8% in 1970–1974, and Google led all organizations in 2020–2024.
  3. Companies supply the most prior-work hires at 7 of the 9 frontier AI labsIncludes Allen Institute for AI (Ai2). Our estimates of the organizations where 550 people worked directly before joining 9 frontier AI labs across 620 career transitions, comparing the main feeder organizations per lab.
  4. 72% of AI researchers with top works connect into a single 2,592-person co-author groupIncludes Allen Institute for AI (Ai2). We map the co-author network across 3,593 people shaping AI and 14,977 most-cited papers: 2,592 (72%) form one connected group with a median distance of 5 steps, led in 2-step reach by Yoshua Bengio (570) and in shortest-path bridging by Yoshua Bengio (6.7%).

Colleagues

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

Sources and credits

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

  1. Shashank Gupta | alphaXiv alphaxiv.org. Checked 20 October 2026.
  2. [2303.17651] Self-Refine: Iterative Refinement with Self-Feedback arxiv.org. Checked 20 October 2026.
  3. [2311.04892] Bias Runs Deep: Implicit Reasoning Biases in Persona-Assigned LLMs arxiv.org. Checked 20 October 2026.
  4. Persona-Bias - Allen Institute of AI allenai.github.io. Checked 20 October 2026.
  5. Best Paper Awards - ACL 2024 2024.aclweb.org. Checked 20 October 2026.
  6. Best Papers - EMNLP 2024 2024.emnlp.org. Checked 20 October 2026.
  7. [2407.18901] AppWorld: A Controllable World of Apps and People for Benchmarking Interactive Coding Agents arxiv.org. Checked 20 October 2026.
  8. [2409.07440] SUPER: Evaluating Agents on Setting Up and Executing Tasks from Research Repositories arxiv.org. Checked 20 October 2026.
  9. ICLR Oral LLM-SR: Scientific Equation Discovery via Programming with Large Language Models iclr.cc. Checked 20 October 2026.
  10. [2512.13961] Olmo 3 arxiv.org. Checked 20 October 2026.
  11. Shashank Gupta shashankgupta.info. Checked 20 October 2026.
  12. Team Directory | Ai2 allenai.org. Checked 20 October 2026.