Pranoy Panda

I am an AI Researcher at Fujitsu Research (India). My current research interests are LLM-guided optimization, open-ended discovery, and evaluation of single-LLM and multi-LLM systems. I am also interested in sequential decision-making problems (bandits, RL, etc.). I wish to contribute to AI for Science problems in the future!

I completed my Master's (by Research) in computer science at the Indian Institute of Technology (IIT) Hyderabad, and my Bachelor's from the National Institute of Technology (NIT) Rourkela. I was fortunate to be supervised by Prof. Vineeth N Balasubramanian for my Masters Thesis.

I am also grateful for the guidance provided by Prof. Martin Barzyck, Dr. Gaurav Sinha, Dr. Ranjan Mondal, and Dr. Sanchayan Santra during different parts of my research journey.

Links: LinkedIn GitHub CV Google Scholar

Email: 1997[first_name][AT]gmail[DOT]com

News

(Sep '25) Very grateful that our Adaptive LLM Routing work was featured in the "Hot" papers of alphaXiv (https://www.alphaxiv.org/), and was #1 on HackerNews (https://news.ycombinator.com/item?id=45094421). More details here
(Aug '25) Two of our works, Adaptive LLM Routing Under Budget Constraints, and Evaluating Compound AI/LLM Systems through Behaviors, Not Benchmarks, were accepted at EMNLP 2025 (findings)!
(May '24) Our work HOLMES: Hyper-Relational Knowledge Graphs for Multi-Hop Question Answering using LLMs has been accepted for presentation at ACL 2024 [Main] (A* NLP and AI conference). Congratulations to my co-authors, Ankush, Chaitanya, Prof. Pratosh, and Manohar Kaul. This is my first publication from Fujitsu Research India.
(Jan '24) Our work on C2FDrone: Coarse-to-Fine Drone-to-Drone Detection using Vision Transformer Networks has been accepted for presentation at ICRA 2024 (A* Robotics and AI conference). Congratulations to my co-authors Sairam and Prof. Vineeth.
(Dec '23) Our work on Real Time Estimation of Weighted Shapley Values has been accepted for presentation at ICASSP 2024 (highest ranked Signal Processing conference). Congratulations to my co-authors Siddharth and Prof. Vineeth. [Link]
(Oct '23) Our work on Interpretable Model Drift Detection has been accepted for presentation at the CODS-COMAD 2024 Conference (also presented at the 2nd Interpretability for Machine Learning and Healthcare Workshop at ICML 2022.) Congratulations to my co-authors Sai Srinivas, Prof. Vineeth, and Dr. Gaurav Sinha (Microsoft Research) [Link]
(Mar '23) Our work on Feature-Interpretable Real Concept Drift Detection has been accepted for presentation at the Trustworthy ML Workshop at ICLR 2023. Congratulations to my co-authors Sai Srinivas, Prof. Vineeth, and Dr. Gaurav Sinha (Microsoft Research) [Link]
(Apr. '21) Our work on Instance-wise Causal Feature Selection for Model Interpretation has been accepted for presentation at the Causality in Vision Workshop at CVPR 2021. Congratulations to my co-authors Sai Srinivas and Prof. Vineeth. [Link]
(Jan. '21) I will be interning at Amazon Science(India) as an Applied Scientist Intern in the summer of 2021.
(Nov. '20) Our work on Forest Ground Vegetation Cover Estimation has been accepted for poster presentation at the Tackling Climate Change using Machine Learning(TCCML) Workshop at NeurIPS 2020. This is joint work with Prof. Martin Barczyk and Prof. Jen Beverley of the University of Alberta Edmonton. Here is the link to the recorded talk from the workshop.
(Sep. '20) I started my Master's at the CSE department, IIT Hyderabad. I will also be working as Research Assistant in the Machine Learning and Vision Lab in the CSE department.
(Apr. '20) I started my remote internship at the University of Alberta, Edmonton, under the guidance of Prof. Martin Barczyk and Prof. Jen Beverly.

Page updated

Google Sites

Report abuse