Work Experience

Senior Machine Learning Engineer

• Organisation - GuardX
• Period - Currently working here

Senior AI Engineer

• Organisation - IMImobile
• Period - June 2019 to December 2020

NLP Engineer

• Organisation - IMImobile
• Period - December 2017 to June 2019

Research Associate

• Organisation - Abzooba
• Period - August 2015 to November 2017

Education

Bachelor of Technology

• Major - Computer Science and Engineering
• University - Indian Institute of Technology Guwahati
• Period - 2011 and 2015

Projects

Alfred - Intent and Entity recognition engine for IMIbot

Role

Lead Engineer

• Worked with latest open-source libraries and 3rd Party NLP Engines to build the intent recognition engine
• Also responsible for building the deployment scripts and logging for maintainability
• Worked on latest NLP techniques, linguistic rules, feature engineering for Machine Learning Models for
  • Intent Classifier
  • Entity Tagger
  • Wordforms Generation
  • Question Detection

Xpresso - Aspect-based Sentiment Analysis Engine

Role

NLP Engineer

• Worked on the development of aspect-based sentiment analysis and trend detection modules in Xpresso
• Also responsible for the deployment of the Xpresso engine
• Used word-vectors to solve unsupervised NLP problems like
  • Aspect categorization
  • Semantic information retrieval
  • Ontology building
• Worked on various POCs including multitude of
  • domains
    • Tax services
    • Dining & hospitality
    • Government
    • Electronics(laptops and mobiles))
  • and problems
    • Resume scoring
    • Fund recommendations based on chat-bot conversations

Subrogation Prediction

Objective

To find the probability of subrogation in the case using adjuster’s notes

Role

NLP Engineer

• Developed NLP rules that analyze the text notes written by adjusters to extract features for the Machine Learning model to find the probability of subrogation in the case
• NLP techniques used - dependency parsing, POS tagging, lemmatization to build rules for feature extraction
• Also worked with word-vectors to build domain ontology for this domain and identify concepts in a given adjuster note

Email Classifier

Objective

Classify emails from retail merchants and suppliers into 10 categories based on the content of the emails

Role

Lead NLP Engineer

• This was achieved using a rule-based feature extraction which identifies the semantic information in the text of the email. The rules were based on the dependency graph and POS tagging of the sentences
• Responsible for the deployment of the solution in client environment(test and production) starting from building the required environment to the final production deployment
• Was an integral part of the discussions with the client throughout the development phase and also during post-production support. This helped in understanding the requirement better and also fastened the process of implementing the feedback into the system

Identifying accuracy issues in food orders

Objective

To identify accuracy issues in food orders

Role

NLP Engineer

• Data - feedback received by the client(retail food chain) from social media, email and internal feedback data etc.
• Text analysis done using Xpresso
• Used dependency parsing and POS tagging of the feedback text received