Contributing towards efficient methods and scientific machine learning integration into FeelMath product line.Quantum computing for scientific machine learning.

Accelerated nonlinear-model evaluation for efficient Training- C++ implementation of Physics ML workflow

A data driven efficient multi-scale computational tool using Hybrid AI framework - NumProDeep-A data science platform based on FE software architecture.It is an automated Framework to identify ML model built for the HPC. Usage of Hybrid ML technique to transfer material info from CTscan to multiscale simulations of injection molded SFRP as an alternate to FFT based solvers. Interpolation and generalization of workflow for all possible micro-structure variations that occur in the injection molded part as a digital twin to SFRP. Novel model distillation ranking algorithm to automated selection of model evaluations

1.Automated Large scale data preprocessing tool for a Graph Nueral Network Surrogate. 2.Creation of novel vector representation concepts to convert nesting data to graph data structure.3.Automated Parameter selection,Automated data generation,Data extraction and Computer Vision based quality evaluation workflow.

Machine Learning based omni-directional control policy estimation from Multi-Body simulations of Human Arm

RoboICSI-A device to peform ICSI-Intra cyto plasmic sperm injectionMicroN-A micro newton force sensor for cell stiffness evaluation