Devi G Kurup

My role as a research assistant in an NLP research team involves following responsibilities :
1. Web Application development - development of a web application using the Flask framework and MongoDB as the backend. This user interface application facilitates data storage and retrieval, specifically for the collection of research data for research purposes.

2. Data Analysis - Analysis of various datasets for the research project. Extracted data from Twitter based on specific keywords and conducted a thorough analysis of the Twitter dataset extracted as part of a Bot detection research project.

3. NLP System Development - developing an NLP system that utilized the BERT model to automatically classify hate speech in tweets. Currently, I am actively contributing to the ongoing development of an NLP system dedicated to information retrieval. This involves collaborating with a team of researchers and implementing diverse algorithms and techniques such as TF-IDF, Word2Vec, BERT, and GPT. Through experiments and evaluations, I identified the most effective approaches for extracting keywords from extensive text datasets.
4. Research Conference Paper - My responsibility also includes the writing of conference papers. I have submitted a paper to the conference, focusing on the UMAP visualization of Embedding models, and awaiting acceptance for this submission.

Developed a web application using Flask framework and MongoDB as the back end for research purpose. Collaborated with a team of researchers to design and implement a user interface data storage and retrieval system.
Data Analysis on twitter datasets for Bot detection research project. Utilized various data analysis libraries of python to preprocess, clean and analyze the collected data. Extracted meaningful insights and presented findings to the research team.
Developed an NLP system that automatically classifies hate speech on tweets using BERT model.
Currently involved in the development of an NLP system for information retrieval, focusing on keyword extraction and sentence similarity. Collaborating with a team of researchers, designing, and implementing different algorithms and techniques such as TF-IDF, word2vec, BERT, GPT for efficient and accurate information retrieval.
Exploring various NLP techniques including Word2Vec, BERT, GPT and lexicon models to analyze and extract keywords from large text dataset. Conducting experiments and evaluations to identify the most effective approaches for keyword extraction in the context of information retrieval.

Build a machine learning model for predicting whether the shipment of product is being delivered on time or not for an E-commerce website.
Train models using modern ML techniques like Logistic Regression, SVM, Random Forest, XGboost, and Decision Tree with hyperparameter tuning and finalized the model with Random Forest with 85% accuracy.
Real Estate price prediction – Exploratory Data Analysis and train models using Regression techniques like Linear Regression and deployed using Flask.
Email Spam Classification – Classify Email into SPAM or HAM.
Build model using Naïve Bayes Classifier and deployed using tkinter.
Twitter Sentiment Analysis – entity-level sentiment analysis to judge the sentiment of the message.
Using NLP text preprocessing methodologies, derive meaningful opinions from the messages and then calculate the sentiment score and based on the sentiment score classify the nature of the judgment as positive, negative, and neutral using VADER sentiment analysis.
Disease detection system using Deep Learning - An end-to-end solution developed to classify pneumonia disease and handles different types of pneumonia.
The user can upload an X-ray image through a UI interface developed using Flask.
The system shares the result information through a messaging system using Twilio.

Text Analytics using R for measuring transparency in corporate sustainability reporting.
Analyzed data in a qualitative and statistical manner using R and with tools like Provalis – QDA Miner, Simsat and Wordstat softwares.
Analyzed research data, interpret to derive meaningful insights and consolidated reports.
Designed and developed responsive web-based applications using ASP.Net, Javascript, C# about Qatar’s artifacts.
Assisted and developed hybrid mobile applications using Ionic framework for Mobile development course.
Prepare reports on scholarly works and other grant related project activities as per requirement.

Performed as a team lead in various projects.
Plan, test and implement various web-based applications using ASP.NET, JavaScript and C#.
Designed, developed, and implemented Web WorkFlows, significantly improving the organization efficiencies.
Worked as DBA for RDBMS related operations, scheduled database backup and recovery of databases.
Developed and implemented enterprise-wide dashboard using Asp.Net, Oracle and Crystal Reports.
Developed and implemented Web Service and Windows Service for the integration of applications between servers.
Developed Android application using Android Studio, C# and JSON and provide support.
Support the customers with troubleshooting and given guidance and assistance to solve their IT issues while implementing software and how to use the applications.
Prepared Training documents for the software and provide Training to the end-users on software use and application.

Coordinating with various project sites on the progress of work, getting regular reports and other documentation as per the requirement of various clients & projects.
Responsible for the preparation of supplier price analysis.
Preparing Reports, Documents & Spreadsheets.
Developed in house applications for inventory management in Access 2007.

Designed, developed, and implemented software applications according to the technical document given by the project manager.
Developed and implemented enterprise-wide dashboard using Asp.Net, Oracle and Crystal Reports.
Support the customers with troubleshooting and given guidance and assistance to solve their IT issues while implementing software and how to use the applications.
Created SQL Tables, Views, Constraints, Synonyms, Sequences, and managed performance and tuning of SQL queries and fixed the slow running queries in production.
Prepared Training documents for the software.
Provide Training to the end-users on software use and application.

Designed, Coded, Tested and Implemented projects according to the technical document given by the project manager.
Created SQL Tables, Views, Constraints, Synonyms, Sequences, and managed performance and tuning of SQL queries and fixed the slow running queries in production.
Maintaining and supporting the system and preparation of various design and implementation documents.

E-Commerce product shipment delivered on time or not – Build a machine learning model for predicting whether the shipment of product is being delivered on time or not for an E-commerce website. The goal is to study their customers to meet their demands to increase revenue. Train models using modern ML techniques like Logistic Regression, SVM, Random Forest, XGboost, and Decision Tree with hyperparameter tuning and finalized the model with Random Forest with 85% accuracy.
Real Estate price prediction – Exploratory Data Analysis and train models using Regression techniques like Linear Regression and deployed using Flask.
Email Spam Classification – Classify Email into SPAM or HAM. Build model using Naïve Bayes Classifier and deployed using tkinter.
Twitter Sentiment Analysis – This is an entity-level sentiment analysis to judge the sentiment of the message. Using NLP text preprocessing methodologies, derive meaningful opinions from the messages and then calculate the sentiment score and based on the sentiment score classify the nature of the judgment as positive, negative, and neutral using VADER sentiment analysis.
Disease detection system using Deep Learning - An end-to-end solution developed to classify pneumonia disease and handles different types of pneumonia. The solution is user-friendly in that the user can upload an X-ray image through a UI interface developed using Flask. Furthermore, the system shares the result information through a messaging system using Twilio.

Text Analytics using R for measuring transparency in corporate sustainability reporting.
Analyzed data in a qualitative and statistical manner using R and with tools like Provalis – QDA Miner, Simsat and Wordstat softwares.
Analyzed research data, interpret to derive meaningful insights and consolidated reports.
Designed and developed responsive web-based applications using ASP.Net, Javascript, C# about Qatar’s artifacts.
Assisted and developed hybrid mobile applications using Ionic framework for Mobile development course.
Prepare reports on scholarly works and other grant related project activities as per requirement.

Performed as a team lead in various projects.
Plan, test and implement various web-based applications using ASP.NET, JavaScript and C#.
Designed, developed, and implemented Web WorkFlows, significantly improving the organization efficiencies.
Worked as DBA for RDBMS related operations, scheduled database backup and recovery of databases.
Developed and implemented enterprise-wide dashboard using Asp.Net, Oracle and Crystal Reports.
Developed and implemented Web Service and Windows Service for the integration of applications between servers.
Developed Android application using Android Studio, C# and JSON and provide support.
Support the customers with troubleshooting and given guidance and assistance to solve their IT issues while implementing software and how to use the applications.
Prepared Training documents for the software and provide Training to the end-users on software use and application.

Coordinating with various project sites on the progress of work, getting regular reports and other documentation as per the requirement of various clients & projects.
Responsible for the preparation of supplier price analysis.
Preparing Reports, Documents & Spreadsheets.
Developed in house applications for inventory management in Access 2007.

Designed, Coded, Tested and Implemented projects according to the technical document given by the project manager.
Created SQL Tables, Views, Constraints, Synonyms, Sequences, and managed performance and tuning of SQL queries and fixed the slow running queries in production.
Maintaining and supporting the system and preparation of various design and implementation documents.

Developed a web application using Flask framework and MongoDB as the back end for research purpose.
Collaborated with a team of researchers to design and implement a user interface data storage and retrieval system.
Data Analysis on twitter datasets for Bot detection research project.
Utilized various data analysis libraries of python to preprocess, clean and analyze the collected data.
Extracted meaningful insights and presented findings to the research team.
Developed an NLP system that automatically classifies hate speech on tweets using BERT model.
Currently involved in the development of an NLP system for information retrieval, focusing on keyword extraction and sentence similarity.
Collaborating with a team of researchers, designing, and implementing different algorithms and techniques such as TF-IDF, word2vec, BERT, GPT for efficient and accurate information retrieval.
Exploring various NLP techniques including Word2Vec, BERT, GPT and lexicon models to analyze and extract keywords from large text dataset.
Conducting experiments and evaluations to identify the most effective approaches for keyword extraction in the context of information retrieval.