My interest lies in the field of Machine Learning, particularly in deep learning using Artificial Neural Networks, pretty much all my projects are oriented toward this field. Apart from that, I love travelling and I am crazy about dancing. I am currently learning contemporary style of dance.
VMware Software India Pvt Ltd, Bangalore, India | Summer Intern
May 2017 - July 2017
Worked on Firewall anomaly detection using ML that involved designing a Machine
Learning model which can learn the known good practices of firewall rules.
5C Networks, Bangalore, India | Winter Intern
December 2016
Worked on automating the brain tumor (Glioblastomas) segmentation using MRI scans. It involved implementing an input cascaded Convolutional Neural Networks using two pathway architecture in order to detect both local and global features of the tumor.
Unnati data labs, Bangalore, India | Summer Intern
May 2016 - July 2016
Worked on the Algorithmic Music Generation project, that involved developing Artificial Neural Network model to generate music using deep learning in Python. The project was aiming to generate Indian classical music based on the genre chosen by the user without infringing the copyrights. The entire training, testing and generation of music were done using AWS instances.
This project is mainly aiming to generate Indian classical music. Music is mainly an artistic act of inspired creation and is unlike some of the traditional math problems. Music cannot be solved by a simple set of formulae. The most interesting and challenging part is producing unique music without infringing the copyright. The generated music has to sound good, and what sounds good is very subjective and varies from culture to culture. Artificial Neural Network/Deep Learning has a wide range of applications, such as in Image processing, Natural language processing, Time series prediction, etc. In this project we are trying to explore the usage of Artificial Neural Network in the field of art. In order to automate the music generation, the model must be able to remember the learnt features over the longer period of time, this is achieved by a special type of Recurrent Neural Network (RNN) called as LSTM (Long Short Term Memory) network. The entire training, testing and generation of music were done using AWS instances.
Given the set of attributes, the artificial neural network model predicts the presence or absence of heart disease, through collation, analysis, synthesis of data and matching of parameters and characteristic features. Realizing the fact that neural networks take significantly large amount of time for training the model, in this project the training period was reduced by parallelizing the training phase using OpenMP API.
Music is a language that follows no strict rules. It is a form of expression that makes us feel different emotions, sometimes even physically changing us. The goal of this website is to provide an automated Web application that allows a user to browse, store his/her favourite music and post questions over the Internet at any point of time as well as to provide effective
decision support tools. The main aim of this project is to provide recommendations to users in order to guide them through the exploration process. Recommender system forms the core of the project. The system keeps track of songs rated by customers and recommends songs basedon their similarity with other customers. Other features of the website include voice recognition, lyrics translation, displaying the most trending songs. Customers can rate the songs, create a playlist, and upload cover songs.
The popularity of Android OS for mobile is inviting the threats such as malwares. Malware is any program or data which affects the working of a device. Thus, malware detection is the invigorating issue in the computer security. This system identifies the malicious apps affected due to malwares. The permissions given by android apps are used as the dataset. The ID3 algorithm is used to classify the app into benign or malicious. The Admin will take care of new entries of malwares as well as apps in the database. The results are shown as whether the given app is malicious or not.
The system uses a Passive Infrared Radiation (PIR) motion sensor to detect the presence of a person within a certain range. Once detected, it triggers the ultrasonic sensor to record the distance of the intruder, lights up LEDs, sounds an alarm and notifies the authorized user through a phone call and SMS.
Raspberry Pi 3 is used to interface the multiple components using the GPIO module of the Rpi Python library. PIR sensor detects the change in infrared radiation when a warm blooded moving object is in its detection range. According to the change in infrared radiation, there will be a voltage signal generated which is then used to turn on the ultrasonic sensor, sound an alarm and turn on the lighting system (LED). Flask micro web framework was installed in the R-Pi to capture and record the distance when the ultrasonic sensor gets turned on. Thus, this saves power consumption and the memory space of the recording system as the LED and ultrasonic sensor are triggered only when PIR sensor detects an object, thus the system starts recording only when the ultrasonic sensor is turned on. Hence, saving memory space. An Android app has been integrated with the system so that users can monitor activity and get alerts in the form of phone calls and text messages.
ChitChat is used basically for chatting purpose with the Client and Server on local networks. Chatting is a method of using technology to bring people and ideas “together” despite of the geographical barriers. This project demonstrates the mode of communication between client and server using socket programming. The client application runs on one of the user’s PC and the server application runs on the other PC that is on the same network. For the communication to begin, the Client sends a request to the Server with an identification name like host address and the port number on which the Server socket is created, the server responds to the request by identifying the client’s address which is already registered, now both the peers are ready to chat. This application has been designed to work under any LINUX and WINDOWS system with Qt4 designer installed on it, and under common network using TCP/IP protocol.
I was one of the 9 finalists of IET-PATW South Asia level. Present around the World (PATW) is the IET’s (Institute of Engineering and Technology) presentation competition for young engineers and technicians, aged 18 to 30, where the competitors present their idea related to any technological domain. There were 10 other participants from various other colleges belonging to Bangalore Local Network, I being the winner, was qualified for the South-Asia finals.
Here is the journey video of IET-PATW South Asia Finals
Speaker at PyCon Portland, USA, 2017
Gave a talk on the improved version of the Algorithmic Music Geeration project at PyCon 2017, Portland, USA on 19th May, 2017. Being one of the very few student speakers, received an extremely positive response and ideas to work even further on this project.
Gave a talk proposal on the same i.e. "Algorithmic Music Generation" which was among the 25 talk proposals that were accepted for presentation at PyCon India 2016, at New Delhi. Presented on 25th September 2016, received an encouraging response and ideas to work on further.
Here is the link to the proposal that got accepted.
