Keshav Aditya R P

Software Development Engineer (SDE) : Full Stack Web Dev - Platform Independent Mobile App Dev | Machine Learning and Deep Learning
keshavaditya26896@gmail.com

Tenacious worker with a goal to maximise the breadth and depth of my knowledge in the field of Computer Science.
Worked as a Software Developer Engineer (SDE) : Web App (ReactJs + Redux), Platform Independent Mobile App (Flutter) and also implemented back-end services in Ruby and Python for over a year to create production grade products.
Used my free time to work as Freelance Data Analyst for a brief period of time, carry out Research in the field of Bioinformatics by leveraging Machine Learning and Deep Learning.
I've also created several innovative MVP's (Minimum Viable Products) while participating at Hackathons.
Apart from this I'm a basketball player who represented my University Team during my Undergraduate Program.

Core Skill Area :

  • Backend: Ruby on Rails,
  • General Purpose Languages: Python , C++
  • Front end: Dart ( FLutter ), HTML, CSS, Javascript, jQuery, ReactJs + Redux, Twitter Bootstrap
  • Database: MySQL
  • Version Control: Github, Bitbucket
  • Tools: VIM, Docker
  • OS: Ubuntu, MacOS, Windows
  • Cloud: DigitalOcean, AWS

Experience

Freelance Data Analyst

Loan Approval Prediction :

  • Used an initial Loan Approval Prediction data set and made it more suitable for practical circumstances seen in the real world by the means of Synthetic Data Generation.
  • Constructed various data visualisation using the python Seaborn Library to investigate if there were clear seperations between those with low and high credit history and check for other patterns. Experimented with MLP, Adaboost, Random Forest , KNN to find the best model to Predict Loan Approval. Ran PCA to reduce feature size and Heat Maps to check for correlations.
  • Finally made an MVP having a python flask API and a webpage that showed a dynamic Loan Tailored Graph (probability of getting loan Vs +/- values from entered loan amount ) for a user based on his required loan amount. The project was also Dockerized for the purpose of deployment.
  • Language & Frameworks: Python (Flask), Javascript
  • Source Code Management: Gitlabs

Time Series Analysis

  • I was involved in another project to perform Time Series Analysis on a large semi-structured DynamoDB data set to show how long it takes a complete the required documentation to submit a proposal requiring a Loan, handle bottle necks if any and also find the most efficient Loan Representatives.
June 2019 - July 2019

Software Developer

Task I
To develop a Web Application to facilitate aggregation of streaming sources for Movies, TV Shows etc.

  • Language & Frameworks: Javascript (ReactJs)
  • Source Code Management: Bitbucket
Task II
To develop a Platform Independent Mobile Application to facilitate aggregation of streaming sources for Movies, TV Shows etc.

  • Language & Frameworks: Dart (Flutter)
  • Source Code Management: Bitbucket
September 2018 - June 2019

Software Developer Intern

Worked extensively on Movie Review Summarisation employing Text Classification. Performed Part of Speech Tagging to extract Entities from Movie Reviews. Involved in Web Scraping and Data Wrangling to construct the required Dataset.

Benchmarking of Movie Review Summarisation :

  • Experimented with various Models such as Glove, Word2Vec, and FastText
  • Trained the Models to pick the best sentence that resembles a summary sentences that a human would choose
  • Language & Frameworks: Python (NLTK), Ruby on Rails
  • Database: MySQL
  • Source Code Management: Bitbucket

Also instrumental in

  • Setting up Apache Thrift Services, AWS ECS Docker Services
June 2017 - August 2018

Front-end Web Developer Intern

This was my first internship and during my brief period:

  • Responsible for Dynamic Report Generation using Angular Js.
June, 2016

Publication

PromoterPredict: sequence-based modelling of Escherichia coli σ70 promoter strength yields logarithmic dependence between promoter strength and sequence

Bharanikumar R, Premkumar KAR, Palaniappan A. 2018. PromoterPredict: sequence-based modelling of Escherichia coli σ70 promoter strength yields logarithmic dependence between promoter strength and sequence. PeerJ 6:e5862 https://doi.org/10.7717/peerj.5862
We present PromoterPredict, a dynamic multiple regression approach to predict the strength of Escherichia coli promoters binding the σ70 factor of RNA polymerase. σ70 promoters are ubiquitously used in recombinant DNA technology, but characterizing their strength is demanding in terms of both time and money. Using a well-characterized set of promoters, we trained a multivariate linear regression model and found that the log of the promoter strength is significantly linearly associated with a weighted sum of the –10 and –35 sequence profile scores. It was found that the two regions contributed almost equally to the promoter strength. PromoterPredict accepts –10 and –35 hexamer sequences and returns the predicted promoter strength. It is capable of dynamic learning from user-supplied data to refine the model construction and yield more confident estimates of promoter strength.

  • Language & Frameworks: Python (Biopython, Webpy), HTML, CSS, Javascript
  • Source Code Management: Github

Honors & Awards

Smart India Hackathon, National-level Hackathon held by the Central Government of India

RUNNERS-UP (Ministry of Empowerment of Persons with Disabilities)

Objective
Built a Platform Independent Mobile Application to track parking for persons with disabilities. Real Time Parking lot status was tracked using sensors and CNN based image classification of occupied and available parking lots.

  • Language & Frameworks: Dart (Flutter), Python (Tensorflow)
  • Database: Firebase
  • Source Code Management: Github
March 2018

iGEM 2017 Boston, Massachusetts

SILVER MEDAL, Application

Objective
Built a MLP to classify Riboswitches, was exposed to working with genome datasets during this time. Also developed the website to showcase the work done by the whole team.

Currently Continuing Research on Classification of Riboswitches to further performace and accuracy:

  • Experimenting with Keras RNN and Hybrid Models such as a RNN coupled with intial layers of a CNN
  • Language & Frameworks: Python (Keras), HTML, CSS, Javascript
  • Datasets: FASTA and CSV Formats
  • Source Code Management: Github
November 2017

SRM ENIGMA V2.0, National Level Hackathon Conducted By SRM University

WINNERS

Objective
Built an Application that aggregated various Government Welfare Schemes under one roof. It finds the best scheme for an individual based on certain field parameters.

Reasons for choosing this Idea

  • Highly dificult for the common man to browse through the internet to find the right governemnt scheme for himself.
April 2017

Student Research Day, Sri Venkateswara College of Engineering

RUNNERS-UP

Objective
Used Multivariate Linear Regression to predict the strength of sigma seventy promoters.

To accomplish this

  • The biopython and sklearn python libraries were used for implementation.
  • Training and testing data set consisted of well characterisedanderson promoters designed at the anderson lab in UC Berkeley
March 2017

Certifications

Machine Learning

Stanford | ONLINE
Associate Professor Andrew Ng, Computer Science Department, Stanford


I would like to take this oppurtunity to thank Professor Andrew Ng, Head Teaching Assistant - Kian Katanforoosh & Teaching Assistant - Younes Bensouda Mourri, supporting cast at deeplearning.ai and Stanford for their efforts in making this a wondeful learning experience and providing us with us quality learning material.

Education

Stony Brook University, New York

M.S Computer Science
Courses : Distributed Systems, Theory of Database Systems, Foundation of Human Computer, Analysis of Algorithms
2019 - 2021

Anna University,   Sri Venkateswara College of Engineering

B.E Computer Science and Engineering
CGPA: 7.82
Courses : Software Engineering, Operating Systems, Machine Learning, Deep Learning, Computer Network
2014 - 2018

Bharatiya Vidya Bhavan's Rajaji Vidyashram


Class 12 (2013 - 2014): 88%

Class 10 (2011 - 2012):8.8

2001 - 2014

Projects

Riboswitche Classifier

Objective
Built a Bioinformatic tool using Machine Learning and Deep Learning models to classify Riboswitches without featureengineering the data. Current version of the model (RNN) can classify 32 classes of Riboswitches with an accuracy of99% and favourable F1 Scor

  • Language & Frameworks: Python (sklearn, Tensorflow - Keras)
  • Datasets: CSV Format
  • Source Code Management: Github
2019

Cluster NBA Draft Prospects

Objective
Currently building a model to cluster similar National Basketball Association (NBA) Draft prospects based on college and draft-combine statistics. Looking to collect data related to players’ background and their mentality.

  • Language & Frameworks: Python (sklearn)
  • Datasets: CSV Format
  • Source Code Management: Github
2018

Smart Security System

Objective
To Build an economically feasible Smart Security System for my household which uses a Siamese Network for Face Recognition to run on a basic microcontroller.

  • Language & Frameworks: Python (Keras)
  • Database: MYSQL
  • Source Code Management: Github
2018

Motorcycle Collision Control System

Objective
Worked on an Collision Control System which sends an alert message to an emergency contact with the GPS coordinates of the accident.

  • Language & Frameworks: Python
  • Hardware: Raspberry Pi, GPS sensor, Accelerometer
2017

Movie Blockbuster Predictor (SAP Hackathon)

Objective
Built a MVP to predict if a movie would be a Blockbuster or not using a Deep Neural Network for classification. Took into account the movie's "story" which proved to be an important parameter

How would this is useful

  • Useful to cutomers as they would know if it would be worth booking tickets for the first day show !
  • Help Cinema Theater Owners gauge which movies would more likely bring them the most returns on their investments
2017

Hatch Hackathon

Objective
Built a platform for doctors to interact with patients.Focusedon digital medical prescriptions which had multiple benefits, viz. –Can ease congestion at the Hospital Pharmacy, can save paper used for prescriptions, and can have a digital trail of all medications prescribed.

2017

iGEM 2016 Website

Objective
To showcase the work put in by the iGEM Team of Sri Venkateswara College of Engineering

2016

IIT Madras Hackathon

Objective
This was my first hackathon. Built a e-commerce price comparison platform based on Web-Scraping

2016

Co-Curricular & Extra-Curricular

Speaker for a Web Development Workshop

Objective
A Web Development workshop by Sri Venkateswara College of Engineering ACM Student Chapter during the Sri Venkateswara College of Engineering Computer Science Symposium Interrupt in collaboration with ACE. The workshop focussed on the LAMP Stack. The various students attending the workshop were asked to setup apache server, a database and were taught HTML, CSS, Javascript, Php, Sql through short crisp examples. The workshop also touched various other concepts in a more theory oriented way such as Ajax for asynchronous Server Requests, hosting/deployment of a website and various other languages currently being used in the industry. To put all what they had learnt to work a demo Resume Website was prepared to demonstrate the interaction between the Frontend and Backend of Web Development.

September, 2017

Executive Member SVCE ACM Chapters

Objective
Being a Member of the Web Team I was responsible for maintaining the ACM - SVCE Students Chapter website and developing new ones along with the team for all ACM - SVCE Students Chapter based Events. APart from this I was also a Event Coordinator for a BioHackathon that took place at SVCE where ACM - SVCE Students Chapter and the iGEM Team collaborated together.

2017 - 2018

Sports Secretary, Student Association at SVCE

Objective
Being a Member of the Student Association at SVCE I was incharge of conducting various events at a intra-college and inter-college level. Being the Sports Secretary I also took the collective opinon of issues regarding the betterment of sport infrastructure and Training. On such instance was that there were only a few athletes who wanted to participate in the Best Physique Contest, due to this it was never included to our Annual Sports Day Festives. I felt that the few who were diligent should not miss out on the oppurtunity and therefore urged for the inclusion of Best Physique and ensured it was.

  • I played basketball for my college team throughout my undergraduate program and captained the team during my final year.
  • The team placed on numerous occasions.
  • I have availed the Sport Scholarship for a span of 3 years
2016 - 2017