I am a Ph.D. student at the University of Trento, working with Bruno Lepri at Fondazione Bruno Kessler (FBK). Previously, I worked at TRDDC, IIIT Hyderabad, Inria, and Columbia University.
My research interests lie in the cognitive science of language models, metacognition and curiosity-driven exploration, pragmatics and strategic communication, and multi-agent AI safety. If you’re interested in my work, feel free to contact me.
🔥 News
- 2026.10: Our preprint Who Else Is In? is available on arXiv.
- 2026.01: I joined the Cooperative AI Foundation as a Research Fellow.
- 2025.11: I started my PhD at the University of Trento, advised by Bruno Lepri.
- 2025: 🎉 Our paper on LLM response sampling received an ACL Best Paper Award!
📝 Selected Publications
Google Scholar ↗Who Else Is In? Multi-Agent Collusion Initiation via Strategic Vagueness [Preprint · 2026]
Humans monitor learning progress in curiosity-driven exploration [Nature Communications · 2021]
A VTA GABAergic computational model of dissociated reward prediction error computation in classical conditioning [Neurobiology of Learning and Memory · 2022]
* Equal contribution / co-first authors.
Additional publications & workshop papers
Multi-State-Action Tokenisation in Decision Transformers for Multi-Discrete Action Spaces.
Perusha Moodley, Pramod Kaushik, Deepak Thambi, Matthew Trovinger, Prudhvi Paruchuri, Xiaoxiao Hong, Benjamin Rosman.
IJCNN, 2024.
EARL: Early Intent Recognition in GUI Tasks Using Theory of Mind.
S. V. Pawar, B. Pedapudi, Pramod Kaushik, S. Sivaprasad, M. Fritz, S. Karande.
ICML Workshop on Computer Use Agents, 2025.
LLM Sample: Part average, part ideal.
Pramod Kaushik, Sarath Sivaprasad, Sahar Abdelnabi, Mario Fritz.
ICML Workshop on Large Language Models and Cognition, 2024.
Momentum Boosted Episodic Memory for Improving Learning in Long-Tailed RL Environments.
Pramod Kaushik, Dolton Fernandes, Harsh Shukla, Surampudi Bapi Raju.
NeurIPS DeepRL Workshop, 2022.
Emergence of Direction Selectivity and Motion Strength in Dot Motion Task Through Deep Reinforcement Learning Networks.
Dolton Fernandes, Pramod Kaushik, Surampudi Bapi Raju.
Conference on Cognitive Computational Neuroscience, 2022.
A Conservative Q-Learning approach for handling distributional shift in sepsis treatment strategies.
Pramod Kaushik, Sneha Kummetha, Perusha Moodley, Surampudi Bapi Raju.
NeurIPS Workshop on Bridging the Gap: from Machine Learning Research to Clinical Practice, 2021.
📖 Education
2025–present
University of Trento
PhD in Information Engineering and Computer Science
2017
IIIT Hyderabad
MS by Research in Computer Science
Working Experience
- 2025.11 – present, PhD Student at the University of Trento and FBK, Trento.
- 2026.01 – 2026.04, Research Fellow at the Cooperative AI Foundation, South Africa.
- 2022.10 – 2025.10, Researcher at TRDDC, India.
- 2021.04 – 2022.09, Research Associate at the Cognitive Science Lab, IIIT Hyderabad, India.
- 2018.10 – 2020.10, Researcher at Inria Mnemosyne and Inria Flowers, Bordeaux.
- 2017.10 – 2018.02, Research Assistant at the Kavli Institute for Brain Science, Columbia University, USA.
- 2016.06 – 2016.12, Research Intern at Inria Mnemosyne, Bordeaux.
🎖 Honors and Awards
2026
Cooperative AI Research Fellowship
2025
ACL Best Paper Award · Co-first author
2025
ELLIS PhD Fellowship
Get in touch
For research conversations and collaboration:
pkaushik@fbk.eu