Portfolio

Personal Statement

I have had the privilege to work on some of the interesting challenges that can be presented to an Engineer from a vast array of application domains. I have used my skills in Machine Learning, Evolutionary Algorithms, Data Analysis, Signal Processing, High-Performance Computing, Embedded Programming, and Effective Visualization to find solutions for problems in extremely disparate domains; which include Big Data, Crowdsensing, Medical Information Processing, Music Information Retrieval, Simulation-based Learning, and many more. Naturally, in the process, I have worked with some of the most brilliant and curious students, academics, and business personnel.

Sharwin Bobde 27 December, 2020

Education

Delft University of Technology (TU Delft)

Masters in Computer Science, with Mastertrack in Data Science and Technology
August 2019 - Present
CGPA: 8.08 (most recent)
Delft, The Netherlands

MIT College of Engineering

Bachelors in Engineering - Information Technology
July 2015 - July 2019
CGPA: 8.33
Pune, India

Experience

Student Assistant: TU Delft

Jul 2020 – Dec 2020
Helping the TROMPA project which is supported by the European Union’s Horizon 2020 research and innovation programme H2020-EU.3.6.3.1. Creating a tool for comparing different types of models used for Music Object Recognition

Student Mentor: TU Delft

Aug 2020 – Jan 2021
Student Mentor for incoming Masters students during COVID19 circumstances at Technische Universiteit Delft

Student Intern, Persistent Systems

Sep 2017 - Jun 2018
Pune, India
A project aimed to create an autonomous drone. Used Unity Game Engine to create an environment to train a Neural Network using Unity ml-agents. Also worked with hardware such as NVIDIA Jetson-TX2 and Zed depth camera.

Volunteering

Chairman, IET Campus Chapter

Jan 2018 - Jun 2019
Pune, India
Headed the IET On-Campus Student Chapter. Responsibilities included deciding, hosting and managing IET sponsored and IET flagship events.

Certifications

Java SE 8 Oracle Certified Associate (OCA)

Publications

1. Courier: Real-Time Optimal Batch Size Prediction for Latency SLOs in BigDL Number, 2021 (accepted but not published)
2. Cognitive Depression Detection Methodology Using EEG Signal Analysis, 2018
3. Different Parameters for Computational Depression Analysis: A Review, 2018
4. An Advanced Design for Depression Analysis through EEG Signal, 2017

You can also check out my Google Scholar page:

Major Projects and Demonstrated Skills

Intent-based Music Recommendation using Evolutionary Algorithms (Masters’ Thesis)

Nov 2020 - Present
Using Evolutionary Algorithms to solve the multiobjective optimization problem of recommender systems. Here, specifically a large scale music listening histories dataset. Currently in progress.

Music genre classification analysis

June 2020
Project for CS4065 at TU Delft. Analysed the problems with the task of music genre classification and large scale datasets in terms of their annotation and collection issues. Used a Convolutional Neural Network to understand the effect of data sampling techniques on classification strength.

One-Shot learning using Siamese Networks

Mar 2020
Project for CS4240 at TU Delft. Reproduced the original project and improved the training process using cyclic learning rate.

Stance Detection for News Articles

Mar 2020
Project for IN4325 at TU Delft. Using Natural Language Processing (NLP) techniques to extract features from articles and exploiting Machine Learning to classify articles as for, against or observing.

Lung Volume Visualization for 4D Lung Data

Dec 2019
Visualization technique to represent lung air-volume movement in a breathing cycle. The practitioner can identify anomalies, tumours, physical deformities and their movement in time.

Sentiment-based Music Recommendation using Machine Learning and Music Information Retrieval.

2018-2019 (Bachelor’s Project)
A recommendation system that dynamically senses the user’s sentiment based on activity and recommends music. Music Information Retrieval is used to extract quantitative features from the music database. User’s activity is mapped to the type of music they like to listen to; this includes the estimated depression status (HAMD)

Persistent Systems: Drone Project

Sep 2017 - Jun 2018
A project aimed to create an autonomous drone. Used Unity Game Engine to create an environment to train a Neural Network using Unity ml-agents. Also worked with hardware such as NVIDIA Jetson-TX2 and Zed depth camera.

Novel Hand Gesture Recognition System using a Sensor Integrated Hand Glove

Nov 2017 - June 2019
A rubber glove with PVDF sensors used for recognizing motion (patent pending). PCB designed in Proteus, assembled with Microcontroller, Wi-Fi module, Analog MUX, voltage regulators, FFC connectors. A wirelessly connected computer runs an app with an interface for classifiers in R. Filed patent 201821020983 A.

Depression Detection and Scaling System

Aug 2017 - Dec 2017
An R code which accepts a person’s EEG data and performs Signal Processing and feature extraction to analyse depression.

Contact

One can always reach me by mail.
For any research-related queries please send a mail to s.p.bobde@student.tudelft.nl.
For any other queries please send a mail to bsharwin@gmail.com.