Sie sind auf Seite 1von 1

Last Updated on 17th February 2020

Anurag Vij
+91 - 7002186709 | vijanurag264@gmail.com

EDUCATION EXPERIENCE
IIT GUWAHATI UNITED HEALTH GROUP (OPTUM) | DATA SCIENTIST
BTECH IN MECHANICAL Nov'18 - Present | Gurugram, Haryana
ENGINEERING • Developed an inhouse Scalable Claims Adjudication Analytics Engine for one of
July 2017 | Guwahati,AS the LOBs of UHG. Tools used: Pyspark(for ETL),D3.js(for visualizations),
Python (for backend) and Markovify and Sklearn ( for modelling
transitions).Savings estimated of about $16M for the company.
LINKS • OCR solution developed for Payments Team to extract few important details
Github:// anuragvij264 from scanned cheque images. Solution currently deployed in production and
LinkedIn:// anuragvij264 being used by Global Payments Team of UHG
Twitter:// @anuragvij264 • FAQ chatbot using a pretrained model named BERT(bert as a service). For
every query,the most similar question was returned as a response.
COURSEWORK TATRAS | DATA SCIENTIST
GRADUATE July 2017 – Oct' 2018 | New Delhi
Deep Learning Specialization • Retrieval Based Chatbot: Built a retrieval based chatbot using siamese
Convolutional Neural Networks for visual architecture to calculate the similarity between contexts and given responses in
recognition the database.
Bayesian Machine Learning • News Recommender system: A content based news recommender system
Machine Learning using the news corpus dataset built.Techniques used to featurize data were
TfIdf, topic modelling and word embeddings.
SKILLS • Automated reports generation system: Developed an application for a client in
Education domain to generate automated reports based on some data which
PROGRAMMING user would provide.Tools used: Python(backend), celery(for async processing),
Expert MongoDB
Python • Pytorch • Tensorflow
• Flask • Sklearn HANYANG UNIVERSITY | RESEARCH INTERN
May 2016 – Aug 2016 | Seoul, SK
Intermediate:
Docker • PySpark • MySQL • Developed an extension of Gustafson Kessel clustering algorithm to Type-2
• MongoDB interval fuzzy sets.
• Used our algorithms on various applications like image segmentation and
Familiar: obtained better results
Apache Kafka • Apache Superset •
Hadoop
SELECT PROJECTS
AWARDS SENTIMENT ANALYSIS OF HINDI-ENGLISH MIXED TEXT |
SEMEVAL 2020 SENTIMIX TASK | GITHUB
• Achieved a place in top 2 in EXL
Excellence Quotient 2016 among 418 • Given around 15K tweets in Hinglish, created a sentiment analysis
other teams from 8 IITs model using Bidirectional-LSTM and self attention.
• Used MUSE to align the word vectors of Hindi and English.
• Secured All India Rank 2426 • Also trained a transliteration model using to get Hinglish to Hindi word
(GEN)among 0.15 million candidates
mappings using (IIT B)parallel data.
appearing for the test
BUILT A CLONE OF REDIS USING PYTHON GITHUB
• Implemented a complete server and client in python
• Implemented Redis protocol parser and MSET,GSET commands in
python

QUESTION ANSWERING SYSTEM GITHUB


• Implemented a Question Answering using pretrained DistilBert Model.
• Currently working to extend this to Indian languages.

Das könnte Ihnen auch gefallen