I am a postdoctoral researcher at Carnegie Mellon University, working under the supervision of Prof. Gauri Joshi. I completed my PhD at EPFL advised by Prof. Anne-Marie Kermarrec. Before that, I obtained my bachelors in Computer Science & Engineering at the Indian Institute of Technology Tirupati.

My research interests lie broadly in efficient machine learning systems for training and inference. During my PhD, I worked on the challenges of communication efficiency, robustness, privacy, and unlearning in federated and decentralized learning. My current research explores flexible model merging and scalable query routing across thousands of LoRA adapters for multi-task deployments.

Selected Publications

Conference Papers

Robust Federated Inference [paper][code]
A. Dhasade, S. Farhadkhani, R. Guerraoui, N. Gupta, M. Jacovella, A. Kermarrec, R. Pinot. ICLR 2026

Navigating the Accuracy-Size Trade-Off with Flexible Model Merging [paper][code]
A. Dhasade, D. Jhunjhunwala, M. Vujasinovic, G. Joshi, A. Kermarrec. ICLR 2026

Revisiting Ensembling in One-Shot Federated Learning [paper][code]
Y. Allouah, A. Dhasade, R. Guerraoui, N. Gupta, A. Kermarrec, R. Pinot, R. Pires, R. Sharma. NeurIPS 2024

QuickDrop: Efficient Federated Unlearning via Synthetic Data Generation [paper][code]
A. Dhasade, Y. Ding, S. Guo, A. Kermarrec, M. de Vos, L. Wu. Middleware 2024

Fairness Auditing with Multi-Agent Collaboration [paper][code]
M. de Vos, A. Dhasade, J. Bourrée, A. Kermarrec, E. Merrer, B. Rottembourg, G. Tredan. ECAI 2024

TEE-based Decentralized Recommender Systems: The Raw Data Sharing Redemption [paper][code]
A. Dhasade, N. Dresevic, A. Kermarrec, R. Pires. IPDPS 2022

Parallel and Scalable Precise Clustering [paper][code]
S. Byma, A. Dhasade, A. Altenhoff, C. Dessimoz, J. Larus. PACT 2020

Workshops

Practical Federated Learning without a Server [paper][code]
A. Dhasade, A. Kermarrec, E. Lavoie, J. Pouwelse, R. Sharma, M. de Vos. EuroMLSys 2025

Harnessing Increased Client Participation with Cohort-Parallel Federated Learning [paper][code]
A. Dhasade, A. Kermarrec, T. Nguyen, R. Pires, M. de Vos. EuroMLSys 2025

Decentralized Learning Made Easy with DecentralizePy [paper][code]
A. Dhasade, A. Kermarrec, R. Pires, R. Sharma, M. Vujasinoic. EuroMLSys 2023

Awards

I was awarded the President’s gold medal for the best academic performance across all branches at IIT Tirupati for my CGPA of 9.75/10.0. I was also awarded the Institute gold medal for standing first in the Department of Computer Science & Engineering over all four years.

Please find a short clip of the award here.