👋🏼 Hello there, I’m Satish!

👨🏻‍💻 I am a PhD student and a Graduate Research Assistant at the Penn State University, USA.

🔬 My primary research interest is in machine learning, distributed learning and natural language processing. Specifically, I am interested in continual/lifelong learning, federated learning and privacy-preserving machine learning. I try to solve the problems through optimizational point-of-view as well as using Bayesian learning techniques.

Humans are naturally capable of lifelong learning and can retain knowledge over time. However, providing computer algorithms with similar cognitive capabilities remains a challenging yet very compelling task. Through my research, I want to understand how to develop efficient yet adaptive lifelong learners which are good at handling the sequential task dynamics. Currently, I study this through the lens of model parameter optimization and probabilistic modelling with emphasis on efficient memory schemes for replay memory. In future, I would like to study how to make transformers and self-attention mechanisms more efficient and adaptive to evolving task dynamics

I am also interested in building distributed and privacy-preserving machine learning and deep learning models through collaborative (federated) learning . Recently, growing data privacy concerns have introduced significant obstacles in developing safe ML models for sensitive applications. I aim to develop federated learning algorithms for safe, reliable and robust ML applications.

News!!

[Aug 2025]: I am joining as a PhD student at the College of Information Sciences and Technology, Penn State (IST) under Prof. Vasant Honavar.

[Jan 2025]: Our work titled “On the Convergence of Continual Federated Learning Using Incrementally Aggregated Gradients” has been accepted (poster) at AISTATS 2025. I am grateful to my co-authors Nazreen Shah and Dr. Ranjitha Prasad. The arxiv version of this paper is available here: arXiv.

[Jan 2025]: We won the “Best Poster Paper Award” at the 26th International Conference on Distributed Computing and Networking (ICDCN), IIT Hyderabad, India. I also presented our work as a lightning talk and poster at the conference. See pics

[Dec 2024]: I gave a talk on federated learning at the CODS COMAD 2024, IIT Jodhpur, India.

[Nov 2024]: My work on the convergence study of continual federated learning is available on arXiv.

[Oct 2024]: My paper on continual federated learning is accepted at CODS COMAD 2024 under the Young Faculty & Researchers’ Track.