Working on Vertically partitioned Federated Learning (VFL) for deep space habitats.

• Developed a decentralized optimization framework for solving large-scale Mixed Integer power system optimization problems.• Successfully used the framework to solve the sensor-data-driven generator maintenance problem coupled with the unit commitment and dispatch problems in a decentralized manner.•. Developed decentralized, Blockchain based algorithms for cyber attack detection for large scale power networks.• Integrated differential privacy concepts in conventional Unit Commitment problems for protection of local generator oriented sensor data.

• Awarded by the Sam Nunn School of International Affairs, Georgia Institute of Technology for research in areas pertinent to national security.• Conducted research on application of decentralized blockchain and machine learning for cyber security deterrence and preparedness with respect to the U.S. power grid.•. Delivered a talk entitled Conflict Reduction and Deterrence in the Power Grid of the Future: A Cyber Security Perspective at the Special Operations Command Centre (SOCOM) of the United StatesDepartment of Defense in Tampa FL

• Developed a scalable, blockchain driven Federated Learning mechanism capable of functionning in a fully decentralized, aggregator free fashion. • Implemented an agile Solidity based SmartContract for decentralized model sharing over a private Ethereum blockchain network.•. Framework was applied towards taxi ride dispatching and improving driver decisions leading to a revenue increase of upto 30%.•. Research culminated in a paper that was presented at IEEE Blockchain 2020

• Developed ACHILES, a distributed memory Asynchronous Iterative Linear Solver based on Message Passing Interface(MPI) and Remote Memory Access (RMA) semantics.• ACHILES is a C++ based High Performance Computing (HPC) solver for scalable asynchronous computation for linear problems.• ACHILES outperformed synchronous versions on 2D finite difference problems using overlapping Jacobi Schwarz.• ACHILES matrix vector products outperformed the synchronous version in Trillinos on a number of large sparse matrices.

• Developed and implemented INDIGO, a decentralized optimization framework based in C, for solving convex seismic tomography problems using distributed gossip protocols.• INDIGO was evaluated on a hardware platform consisting of BeagleBone Black interfaced with XBee Pro S3B radio as well a simulation platform over the CORE network emulator.• Developed an iterative learning algorithm for identifying regions in the network having high latency, high delay in large scale Wireless Mesh Networks culminating in a paper accepted at IEEE WoWMoM 2015.• Developed radio communication software module for hardware nodes consisting of MSP430 microcontroller and an XBee Pro S3B radio, participated in deployment of 20 such nodes on a volcano in Llaima, Araucania, Chile in January 2015.

Developed a software testing framework for Reliability and Availability testing of the NetWorker server.

Developed an Online Exam Manager using ASP.NET and C# as scripting language.