Hi there 👋 ... Welcome to my personal git repo. I hope you'll learn at least something here. You can find my codes in various different categories including the following:

• Research

  • Research in computational condensed matter physics. Check my webstite [email protected]
  • Introduction and real applications of Density Functioal Theory (DFT) and embedded Dynamical Mean-Field Theory (eDMFT) calculations on quantum materials.
  • Python TightBinding models for various Topological phases of real materials.

• Programming

  • python for physics research as well as data science.
  • Linux and bash scripting language
  • grep, egrep, sed, awk and vi editor

• Machine Learning and Deep Learning

  • Implementation from scratch (only using numpy)
  • Real world application of various ML methods.
  • Natural language processing
  • Computer Vision
  • Productionization: Deploying your model as a web app using Falsk and Streamlit.

• Quantum Computing and Quantum Machine Learning

  • Basic quantum computing algorithms and it's implementation using qiskit.
  • Introduction to some quantum machine learning algorithms.
  • Check here for code: https://github.com/quantumyatra/

• Finalist of the Erdos Data Science Bootcamp 2021.
• Scraped >2M Reddit posts from Reddit and >72K unique tweets mentioning GameStop.
• Wrangled the features to extract intrinsic properties of the posts along with external information.
• Built predictive models using word vectorization and probability calibration on an imbalanced (5:95) dataset.
• Deployed the model via Heroku with the option to choose between aggressive (high recall) and defensive (high precision) strategies.
• Website: (https://gristock.herokuapp.com/)

• Project work for NLP-Pilot Bootcamp organized by Erdos Institute.
• Processed the highly imbalanced data set containing Wikipedia comments posted online.
• Used Logistic Regression, Random Forrest, XGBoost, and various Neural Network architecture for the multilabel multiclass classification problem.
• Developed a web app and deployed it online (https://toxcom.herokuapp.com/) for any user to find whether an online comment is toxic or not as well as to distinguish what kind of toxic label the comment has.

• Cleaned the real-time solar wind data and imputed it using forward fill and interpolation methods to predict DST index.
• Performed time-series analysis using ARIMA model as well as implemented LSTM model in TensorFlow-2 framework.
• Ranked within the top 8% (group of four) among more than 600 participants in the competition for the LSTM implementation
• Find Code here github.com/khanalg44/MagNet.

If you are into cricket then please check my website https://cric-data.herokuapp.com/ for players' profile, top statistics, analysis in various formats of the game

• Personal webpage: https://khanalg44.github.io/
• GitHub: https://github.com/khanalg44/
• LinkedIn: https://www.linkedin.com/in/ghanashyam-khanal/
• Email: [email protected]
• Rutgers webpage: https://www.physics.rutgers.edu/~khanal/