Awarded with 'Best Intern Demo' at Intern project expo.

Increased portal’s load speed by 16 times by streamlining existing code and configuring the application for AOT build

Designed a data collector to parse data from e-commerce websites and real-time social media APIs like Twitter Firehose.

Improvised NLTK raw text processing to adapt our data feed, increasing the efficiency by ~20%

Achieved 88% classification precision by using a combination of N-Gram POS taggers and Brown Corpus as the training set.

Designed a custom language parser for intent extraction and performed sentimental analysis with 84% accuracy.

The data was visualized in various business-oriented views using the D3.js library.

The system was automated and hosted on AWS EC2 and mLab Data so that the real-time data can be collected, processed and update the views accordingly.