• Train News Relevancy and Categorization classifier using GLOVE and LSTM/GRU.
• Build near-duplicate detection pipeline to identify duplicate news using SimHash.
• Identify entities like person name, geo-location, organization etc. (NER) in client specific news.
• Design REST API functions using Flask for connecting frontend with backend.
• Performing CRUD operations on unstructured news data using MongoDB and Elasticsearch.
• Extract morphological patterns from news to generate automatic alerts, news digest and reports.

• Developed multi-document topic identification framework for identifying topics being discussed about each client in news.
• Implemented extractive summarizer to generate summaries for news articles.

“Sector Neutral Stock Classification Model for maximizing returns”

• Develop models for portfolio risk assesment.
• Identify feature for all the sectors, thus resulting in a model that is Universal for all sector.

Deep Learning International Fellow

• International Fellow mentored By Dr. Jeremy Howard.

Gender Prediction using Multimodal Feature Engineering

• Gender prediciton of Twitter Users based on Image(Vision), Tweets(Context), Profile Description(Personality) and Name(Ethnicity/Nationality).
• Language predictor using stopwords matching.

Brand Identificaiton based on Twitter User's' Profile Image using Convolution Neural Networks

• Extracted features from images using VGG-16 model,a 16-layered convolutional architecture with transfer learning.
• Used AWS EC2, S3, and p2.xlarge services.

How Diverse Users and Activities Trigger Connective Action via Social Media: Lessons from the Twitter Hashtag Campaign #ILookLikeAnEngineer

• Temporal and Spatial Analysis.
• Pattern analysis.
  
  1. Participation (Who tweets and shares).
  
  2. Sharing (What messages and resources are tweeted and shared).
  
  3. Attributes of messages (Tone, sentiment, affective and cognitive attributes).
  
  4. Network Analysis( Centralities).

Interactive web application using D3.js to analyze how well United Nation's' Anti Gender Violence Campaigns

• Finding out important actors in driving the campaigns at state, city and country level.
• How well people react to such campaigns.
• Predicted impact factor of campaigns based on retweet counts and favourites.

Started Master's in Data Analytics Engineering

• Deep Learning, Machine learning, Statistics

Graduated with Bachelor's in Computer Science

Supply Chain Analytics Software

• Developed a desktop application (Java,SQL,Python) to track inventories and detect anomalies.
• Developed a three tier architecture for secured user authentication and inventory management.