I am a PhD student at the Scalable Computing Systems Lab, EPFL. Under the supervision of Prof. Anne-Marie Kermarrec, I am working on designing efficient and privacy-preserving systems for Decentralized and Federated Learning. Presently, I am working on Self-Organizing AI as a visiting Ph.D. student at the Camera Culture Group of the MIT Media Lab.
I received my Bachelor’s Degree in Computer Science and Engineering from the School of Computing and Electrical Engineering (SCEE) at Indian Institute of Technology Mandi ranking first in my class. My winter semester 2019-2020 was spent at RWTH Aachen University Germany as part of the Student Exchange Programme, where I attended challenging courses and worked as an Undergraduate Research Assistant with the IT-Security Group and the Theory of Hybrid Systems Group. As part of my final-year Major Technical Project, I worked with Dr. Manas Thakur in the Compilers and Programming Languages Group at IIT Mandi on Automating Loop Parallelization for TornadoVM.
Ph.D. in Computer Science, 2021 - Present
École Polytechnique Fédérale de Lausanne (EPFL)
Student Exchange, WS 2019 - 2020
RWTH Aachen University
B.Tech. in Computer Science and Engineering, 2017 - 2021
Indian Institute of Technology Mandi
Refactored Fixed Data Table 2 Resize and Reorder functionalities into plugins to make the React-JS based library modular, more customizable and maintainable.
Received PPO for the Member-Technical profile on the basis of excellent performance during the internship.
Implemented methods to perform gradient-based attack and adversary transfer on character-based Deep Neural Networks for malicious DGA generated domain names by emulating and inverting the non-differentiable embedding layer.
Used Iterative adversarial training to improve the robustness of the classifier using adversaries generated from the gradient based attacks.
Formulated the scheduling of a freight train in the german railway network as a satisfiability problem in propositional logic and implemented the solution using Z3 Solver.
Optimized the various steps of the problem formulation to reduce the problem blow-up from quadratic to linear and improved the feasibility of the approach in real-life railway network.