Karnesh Jain

Experience

• 

  Mangalore Refinery & Petrochemicals LTD - India

  Jul 2009 - Jul 2012

  Process Engineer

  ▪ Supervised implementation of statistical data reconciliation tool “Production balance” from Honeywell International Inc., India. The project aimed at implementing a real-time database management system. This project helped in increasing the profitability by $0.5M.▪ Created a SQL database for identifying production losses.

• 

  University at Buffalo

  Jan 2013 - Apr 2019

  ▪ Developed a probabilistic Monte Carlo algorithm to study wetting phenomena for enhanced oil recovery. ▪ Developed a high-performance parallel Monte Carlo simulation package in FORTRAN to study phase and interfacial behavior of complex fluids.▪ Created a workflow for running Etomica (Java) and LAMMPS (C++) consecutively to optimize the molecular simulation algorithm for computing wetting properties.▪ Automated data analysis of large data sets (~600 GB) using bash scripts and python libraries to acquire thermodynamic properties. Show less

  • Volunteer Research Assistant

    Sept 2018 - Apr 2019

  • PHD

    Aug 2012 - Sept 2018

  • Teaching Assistant

    Jan 2013 - Dec 2013
• 

  University of North Dakota

  May 2019 - Mar 2021

  Postdoctoral Research Fellow

  ▪ Investigated folding of protein implicated in Parkinson's disease using replica exchange molecular dynamics▪ Built deep learning algorithms (image classification) for improving cancer tumor prediction.▪ Built a python library that enables researchers to perform Monte Carlo simulations.▪ Mentored graduate students on scientific software troubleshooting and data analysis.

• 

  The University of Alabama

  Apr 2021 - now

  Research Software Engineer

  ▪ Developed a deep learning (convolutional neural network) model for multi-label image classification of 2-D materials with 90% accuracy.▪ Developed efficient pipeline to query and visualize from large dataset (4.1 TB) of National Water Model simulation data from Google Cloud Platform that decreases query time by 40%. ▪ Designed and built a web application for displaying soil moisture data from the random forest regression model.▪ Developed GIS web application using Django for interactive visualization of National Snow Model data generated using artificial neural network model and deploying it using multi-Docker container. ▪ Developing agent-based modeling software (NarcoLogic) to model narco-trafficking networks and counterdrug interdiction forces. ▪ Aided in troubleshooting of C++ software package with embedded decision tree-based Python script. ▪ Designed a database (CouchDB) to store materials properties obtained from running Monte Carlo simulations. Show less