Harisyam Manda

I currently work on Preventive Analytics mainly in the following areas: 1) Anomaly Detection in Time Series Signals 2) Remaining Useful Lifetime for NOX Catalysts on Gas Engines 3) Generative Adversarial Networks (GAN'S) and Variational Autoencoders (VAE) for Time Series Data Augmentation 4) Condition Based Monitoring Softwares used are: - PyTorch for doing DL - Software dev with Python and Java - Data Analysis with R - SQL and Cassandra DBs

Root Cause Analysis and Forecasting of Vehicle fleet for durability using Machine Learning My Thesis involves : 1) Exploratory Data Analysis from DAIMLER Truck Data Lake 2) Supervised Learning Algorithms Survey 3) Dealing with highly skewed datasets 4) Developing Robust Logisitic Regression, Robust SVM Classifiers and DL Classifiers for training Data with Label Noise 5) Possibility of performing Correspondence Analysis (CA) and using xgboost ensemble algorithm to find out the feature importances

Development of NVH Big Data Analysis Software with Implementation of Parallel Computing - To create a GUI Post Processor for analyzing the NVH Big data from experiment using Python. - To develop opengl based 3D rendering graphics in the GUI - Creating NVH parameter plots in N>3 Dimensions (Hypersurfaces)

Teaching Assistant for the course Numerical Methods in Mechanical Engineering. - Conducted MATLAB tutorials for the course

Developing an algorithm for determining the solution space and boundary conditions for design curves in Optimization of wheel suspension kinematics. - Development of a TAGUCHI Design algorithm for the creation of solution space - MATLAB Code for the TAGUCHI DOE method to predict the boundary curves - Testing of the MATLAB Code using various suspension data - Implementation of the boundary curves results into Optimization Code

To model the physical non linear behavior of filler reinforced Elastomers using Mullins effect and Dynamic Flocculation Model. - Developed an RVE for filler reinforced elastomers using self avoiding random walk (Advanced Brownian Motion) in ABAQUS using Python - Implemented Dynamic Flocculation Model for the RVE generated above - Extracted the compound E Modul for the Elastomer using material subroutines in ABAQUS.

Project: Noise Optimization of Power Train Components Using ERP Method in Optistruct • Developed a method to do noise optimization using ERP method on various powertrain components. • The Optimization methods used are Topography, Topology, Size and Shape. Combined optimization is also used in the case of Oil Pans • Tools used: Altair Hyperworks, Scripting Language used: Tcl/Tk, Solver Used: Optistruct

FEM, Mechanics Of Materials, Tensor Algebra, Continuum Mechanics, Practical FEM, CFD, Multi Body Dynamics and Computational Solvers (CAE)

CAD/CAM, Heat Transfer, Vehicle and Machine Dynamics, Machine Design, Organic Chemistry