Avatar

Pashupati Gupta

Senior Undergraduate

Indian Institute of Technology Roorkee

About

I'm a final-year undergraduate student at the Indian Institute of Technology Roorkee. My area of interests mainly iclines towards Deep Learning especially in Natural Language Processing. I have demonstrated history of working with several IT industries including tech giants like IBM. I'm skilled in various Deep Learning frameworks and python development along with good management and team work skills.

Interests

  • Deep Learning
  • Dialouge Systems
  • Natural Language Processing

Education

  • B.Tech (Polymer Science & Engineering), 2020

    Indian Institute of Technology Roorkee

Experiences

 
 
 
 
 

Deep Learning Intern

IBM India

May 2019 – Jul 2019 Bengaluru, India
Develop a NLP pipeline which can detect offensive language/text. OLID v1.0 dataset was used for this purpose. Bidirectional Encoder Representations from Transformers (BERT) and XLNet models were tuned on the processed dataset which resulted in a final accuracy of 82% approx.
 
 
 
 
 

Data science Intern

FlyNava Technologies

Nov 2018 – Dec 2018 Bengaluru, India
Developed Python programs to generate Triggers, an Automatic Alert System based on market fluctuations and irregularities, performance variation, events, competitor action, and seasonality. MongoDB database was analyzed by integrating the database with python using PyMongo.

Projects

Check out some of the work done by me

Jarvis

Jarvis is a chatbot specifically created to work as a StackOverflow Assiatant but it can also do some normal chit-chat

Dolores

Dolores doesn't like offensive language. She can very precisely classify a text into offensive/Non-offensive

Estimator

Estimator is a Machine Learning enabled platform to estimate critical chain length of a polymer corresponding to its saturated …

Featured Posts

Building a Text Classification model using BiLSTM

Text classification is one of the fundamental tasks in NLP. Almost every NLP system uses text classification somewhere in its backend. …