Deep Lokhande

Education

Rutgers University

08/2018 - (Expected)07/2021

Electrical and Computer Engineering

Master of Science

University of Pune, Pune, India

07/2013 - 06/2017

Electronics and Telecommunications

Bachelors

Experiences

CS 142 - Data 101

02/2021 - 06/2021

Part Time Lecturer

Rutgers University

EE 366 - Digital Electronics

02/2020 - 06/2020

Teaching Assistant

Rutgers University

Prof. Emina Soljanin's Lab

05/2019 - 08/2019

Research Assistant

Rutgers University

Researched on Quantum Approximate Optimization Algorithm(QAOA) for solving graph based and semi-definite programming problems using NISQ computers.
Implemented the QAOA algorithm for finding the Weighted Max-Cut of a graph using IBM Qiskit.

University of Pune

09/2017 - 05/2018

Research Assistant

University of Pune, India

Researched on unexplored human bio-metrics and security features which can be used in future security system’s with higher security, low cost and system requirements.
Modelled bi-layered security system based on palm print and vein features using low cost DSP’s.

Projects

Data Prediction and Interpolation

Analysed raw data from ManyLab’s project, to predict missing data and extract significant features.
Implemented one-hot encoding & normalization on raw data and programmed Bayesian Network tree using Chow-Liu algorithm for feature dependency interference and data interpolation. Predicted satisfactorily(near 67%) of missing data while eliminating irrelevant features present in data.

Data Analysis and Price Prediction of Airbnb Listings

Exploratory data analysis on Airbnb listings in NYC, to detect hidden trends and patterns.
Applied Linear Regression, Random Forest & Xtreme Gradient Boosting, with cross validation and feature engineering for parameter tuning, resulted in improved Price prediction accuracy with minimized RMSE.

Colorizer for Black & White Image

Modeled a linear regression based learning model in python for converting black and white images to color using training on color and black/white duals of images without semantic classification of objects.
Used stochastic gradient descent optimization to minimize MSE objective function & studied factors beneficial to model performance in learning conditions limited to non-contextual & localized features.

Quantum Error Correcting Codes

Studied techniques used in quantum computers and circuits for error correction and fault tolerance.
Implemented the bit flip and phase flip quantum error correcting code, using IBM’s Qiskit services on IBM’s 5 qubit quantum computer.

Technical Skills

Deep K. Lokhande