Gaurav Devdikar

Developed bash & TCL scripts for automatic pre-processing and scripts for creating input decks for LS DYNA & LS OPT for performing crash simulations & optimization. Designed and implemented custom applications using LLM (GPT-4)+ KG (Knowledge Graphs) in Python Developed Full stack web app chat interface using Plotly Dash

Developed Python implementation for discretization of PDE with triangular elements using Finite Volume Method (FVM) based on reference literature.

Developed of AI/ML based pipeline for condition monitoring of machines using Deep learning architectures. Extracted custom signal processing features based on Wavelet transform to improve ML classifier performance.

1. Enhanced defect mapping for transmission assembly testing by extracting NVH (Noise Vibration Harshness) features using and correlating them with assembly process parameters. 2. To improve confidence over quality of Transmission assembly, proposed a MATLAB/SIMULINK model to correlate gear shift force from functional testing with actual shift force experienced by the driver.

I joined IKSC Knowledge Bridge,Pune training institute focused on Industry related skillsets in the CAE field. I did projects focused on 1D mathematical modeling, Structural FEA, Optimization using Optistruct, and Conjugate Heat Transfer Analysis using Acu Solve. Got hands-on practice on different tools like Hypermesh, Abaqus CAE, Acusolve etc. and Full Vehicle deck preparation for LS-DYNA based simulations

Focus: Finite Element Method (FEM), Multifield Problems, Scientific Computing & Programming with C++, Machine Learning methods, Computer Vision, Data Analytics & Scalable Data Engineering (Topper in the class) Research Project: Constitutive Modeling of Hyperelasticity using Physics informed neural networks Master Thesis: Application of Large Language Models in Automotive Product Development Lifecycle. Collaboration with Fraunhofer SCAI.