Paritosh Ramanan

I specialize in developing high performance tools and applications geared towards scalable decision making for large scale industrial problems. My main area of expertise is in asynchronous and decentralized computational architectures that promise faster solutions with good quality and greater computational efficiency. Asynchronous decentralized methods can be applied to solve problems in areas such as machine learning, power systems and operations research. As a result Im looking at applying my knowledge in asynchronous decentralized computations in emerging areas such as machine learning and blockchain.

As part of my research I heavily use Message Passing Interface(MPI) to orchestrate the asynchronous computational model on High Performance Computing (HPC) clusters. I also have experience in blockchain based decentralized applications (dApps) primarily targeting Federated Learning problems in the field of Machine Learning.

For my PhD research I have developed asynchronous, decentralized solutions to large scale sensor driven Mixed Integer problems for power systems. I developed asynchronous iterative linear solvers as part of my PhD research internship at Sandia. I also designed and implemented a decentralized Federated Machine Learning framework for smart asset management using the Ethereum blockchain based SmartContracts. 

External Advisor
  • Prof. Edmond Chow