LexisNexis, Raleigh, NC, USA

Software Engineer Intern • May, 2019 — present

Search Relevance based Re-ranking Models

  • Created machine learning enabled service to recognize malformed or partially formed legal citations in search query. Success rate of 89%.
  • Partial citation recognition service integrates to core search for Lexis Advance. Java based machine learning frameworks were used.

Wynk Limited, Gurugram, Haryana, India

Senior Software Developer - Backend • Dec, 2016 — Aug, 2018

Machine Learning Initiatives - Wynk Music.

  • Created algorithm, API design and architecture for Learning-to-rank Solr-based auto-search that serves more than 1 million queries/day.
  • Created Search API and algorithm using Solr, Solr-LTR plugin implementing Linear Regression ML model, and extensive caching based microservice architecture that serves 5 million queries/day
  • Created Deep learning (LSTM) model for predictive classifying of native and non-native artist/album complete or partial name inputs.
  • Created Hybrid Recommendation Engine based on matrix factorization (collaborative filtering), cosine similarity and temporal user context.
  • Scaled infrastructure and removed processing bottlenecks for back-end services, databases and data pipeline to handle nearly 1.5 billion events/day.
  • Created architecture and design of Music Backend system implementing microservices in dockerized containers serving 50 million streams/day.
  • Single-handedly managed 4 major components - Analytics Data pipeline, Music Streaming, Search and Discovery, and Auto-suggest.

ANI Technologies Private Limited (Ola Cabs), Bengaluru, Karnataka, India

Senior Software Engineer • Jul, 2015 — Nov, 2016

Algorithm, API, Architecture and Analysis - Ola Share.

  • Designed matching algorithm and scalable backend APIs using extensive caching that serves nearly 0.5 million bookings/day.
  • Single-handedly designed algorithm and services for dynamic predictive pricing and discounting based on probabilistic match.
  • Created predictive algorithm based optimized matching in hot-spot zones resulting in increase of matching path overlap from 29% to 67%.
  • Analyzed vast amount of historical route patterns and demand affinity to create Loose Matching algorithm that caters to 90% of >10 km trips.
  • Promoted within 9 months of experience with only one to be awarded ``5 star'' work performance among 120 teammates.

Axero Solutions LLC, New Delhi, Delhi, India

Software Engineer (Trainee) • May, 2014 — Dec, 2014

Summer internship and part-time web development - Communifire.

  • Developed NUnit test Suite for the code pieces of the software product integrating in Visual Studio.
  • Developed Android app for MCD Delhi for live consumer complaints registration with pictures.
  • Being the best performer during the internship among all other interns, was offered to work part-time as software developer during final year of IIT.
  • Android app development for offline/online updates for Business collaboration software.
  • Added accessibility features to flagship software product to make it compatible for Government websites( access for visually impaired).


North Carolina State University, Raleigh, NC, USA

Master of Science, Computer Science • 2018 — 2020

Undergoing Post Graduation in Computer Science (specializing in Machine Learning). Courses —

  • Automated Learning and Data Analysis
  • Foundations of Software Science
  • Compiler Construction
  • Algorithms for Data Guided Business Intelligence
  • Neural Networks and Deep Learning
  • Aritifical Intelligence

Indian Institute of Technology Delhi, New Delhi, Delhi, India

Bachelor of Technology, Electrical Engineering(Power) • 2011 — 2015

  • Thesis Design of fifth-order boost converter - ANT colony approach. Advisor Prof. M. Veerachary. Published in IEEE ICPS, March 2016 IEEE publication

Member- IIT Delhi Athletics team, Finalist in 3 events- Inter IIT Sports Meet 2012 & 2013, Karakoram House Captain-Basketball


Founder & Primary Developer • Oct 2018 — Oct 2018

Won 2nd Runner-up among 22 teams. Created a highly accurate ensemble of following Models for validating a sequence as a Legal Reference

  • Frequently occurring Regex patterns.
  • Naive Bayes and SVM Classifiers based on attributes derived from Regex patterns and heuristics around sequence of characters.
  • LSTM (Deep Learning) based Sequence Analysis.

Creator • Aug 2018 — Nov 2018

Automatic classification of a commit message as a bug-fix by NLP, SVM and Topic Modeling (LDA). Technologies

  • Python, scikit-learn
  • Support Vector Machine, Natural Language Processing, Latent Dirichlet Allocation

Creator • Sep 2018 — Nov 2018

Classification of a university web page (dataset — WebKb) basis textual content by Deep Learning (LSTM) and NLP. Technologies

  • Python, scikit-learn, Keras, TensorFlow
  • Long Short-term Memory, Natural Language Processing

Creator • Oct 2018 — Nov 2018

Create a compiler that converts Deep Learning models described in Caffe Prototxt to TensorFlow code. Technologies

  • Java, ANTLR



Java • Python • Ruby • PHP

Technologies and Frameworks (Software Development)

Spark • Storm • Kafka • Spring • Spring Boot • Dropwizard • Apache Solr • MySQL • MongoDB • Hazelcast • Redis

Technologies and Frameworks (Data Science)

Spark SQL • Spark ML • Numpy • Pandas • scipy • scikit-learn • Keras • TensorFlow

Areas of Proficiency

Deep Learning • Machine Learning • Computer Architecture • Algorithms • Databases • Scalable Architecture • Data Mining and Analysis


Deep Learning Specialization • Sequence Models • Convolutional Neural Networks • Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization • Structuring Machine Learning Projects • Neural Networks and Deep Learning • JAVA Programming Language • Databases