Shashank Gupta

I'm a researcher at the Allen Institute for AI (Ai2) in Seattle.

My current research centers on developing agentic large language models—systems that can plan, reason, use tools and APIs, write code, and interact with complex environments to accomplish real tasks. I study how training data, objectives, and learning environments shape the emergence of robust and generalizable agentic behavior, and how such capabilities can be measured and controlled in reliable ways.

If you're excited about advancing open agentic LLMs that can reason, plan, and use tools, we're currently hiring interns—apply here or reach out to me directly!

Selected Projects

	OLMo 3 Olmo Team, Allyson Ettinger, Amanda Bertsch, [...], Shashank Gupta, [...], Noah A. Smith, Hannaneh Hajishirzi Core Contributor paper \| code \| data & models \| blog \| X (Twitter)
	LLM-SR: Scientific Equation Discovery via Programming with Large Language Models Parshin Shojaee, Kazem Meidani, Shashank Gupta, Amir Barati Farimani, Chandan K Reddy ICLR 2025 (Oral) paper \| code \| X (Twitter)
	SUPER: Evaluating Agents on Setting Up and Executing Tasks from Research Repositories Ben Bogin, Kejuan Yang, Shashank Gupta, Kyle Richardson, Erin Bransom, Peter Clark, Ashish Sabharwal, Tushar Khot EMNLP 2024 🏆 Outstanding Paper Award 🏆 paper \| code \| dataset \| X (Twitter)
	AppWorld: A Controllable World of Apps and People for Benchmarking Interactive Coding Agents Harsh Trivedi, Tushar Khot, Mareike Hartmann, [...], Shashank Gupta, Ashish Sabharwal, Niranjan Balasubramanian ACL 2024 🏆 Best Resource Paper Award 🏆 paper \| website \| code \| video \| poster \| X (Twitter)
	Bias Runs Deep: Implicit Reasoning Biases in Persona-Assigned LLMs Shashank Gupta, Vaishnavi Shrivastava, Ameet Deshpande, Ashwin Kalyan, Peter Clark, Ashish Sabharwal, Tushar Khot ICLR 2024 paper \| website \| code \| poster \| data \| X (Twitter)
	Self-Refine: Iterative Refinement with Self-Feedback Aman Madaan, Niket Tandon, [...], Shashank Gupta, [...], Peter Clark NeurIPS 2023 paper \| website \| code \| poster \| Forbes
	Sparsely Activated Mixture-of-Experts are Robust Multi-Task Learners Shashank Gupta, Subhabrata Mukherjee, Krishan Subudhi, Eduardo Gonzalez, Damien Jose, Ahmed H. Awadallah, Jianfeng Gao Microsoft AI Journal, 2022 paper
	Knowledge Infused Decoding Ruibo Liu, Guoqing Zheng, Shashank Gupta, Radhika Gaonkar, Chongyang Gao, Soroush Vosoughi, Milad Shokouhi, Ahmed Hassan Awadallah ICLR 2022 paper \| code
	Exploring Low-Cost Transformer Model Compression for Large-Scale Commercial Reply Suggestions Vaishnavi Shrivastava^, Radhika Gaonkar^, Shashank Gupta^, Abhishek Jha equal contribution Microsoft AI Journal, 2021 paper \| blog post
	CogCompNLP: Your Swiss Army Knife for NLP Daniel Khashabi, Mark Sammons, [...], Shashank Gupta, [...], Dan Roth LREC 2018 paper \| code
	Web-scale entity annotation using MapReduce Shashank Gupta, Varun Chandramouli, Soumen Chakrabarti HiPC 2013 paper \| slides

Template credits: Unnat, and Jon