Professional Experience

Lenovo

Cloud Development Engineering Intern

  • Developed a secure authentication and consent app for client access authorization in a connected IoT cloud framework.
  • Extended feature support for handling dynamic user language preference update, addressing internationalization.
  • Successfully implemented a fault tolerant network of interaction in the modular containerized framework, leading to increased security and isolation.
  • Integrated an automated end-to-end test pipeline for stress testing individual modules by mocking user identification.
  • Facilitating transfer of knowledge to global HQ in China for production.
  • Gained exposure to technologies across multiple domains including Docker, Node.js, Express, Pug and batch scripting.

University of Massachusetts Amherst

Graduate Research Intern with Prof. Rene Just

  • Designed and developed a standalone mutation analyzer for the MAJOR mutation framework.
  • Programmed the output interface to summarize the results of the analysis carried on the test suite.
  • Developed examples and integrated the analyzer to the previous release after restructuring and using version control.
  • Automated the multistep process of running the standalone analyzer on its corresponding example.

Nectar Globe Technology Solutions

Software Development Intern

  • Developed a global tweet extractor that accepts multiple parameters as input criteria for searching relevant tweets.
  • Analyzed the stored tweets using ElasticSearch and generated required insights and predictions for market research firms.
  • Developed and designed the front end of an Indian hotel room booking website.

CMC Ltd.

Software Engineer Trainee

  • Trained to develop project modules (C++, Java) as a part of the software development process for product deployment.

Projects

Skill Taxonomy Enhancement for Burning Glass Technologies, Boston advised by Prof. Andrew McCallum

February 2018 - Present

  • Extract unstructured Wikipedia data to run NER model to identify possible new skills that can be added to the taxonomy.
  • Use structured wiki data to analyze relational or hierarchical tree structure of skills to create knowledge graph of skills.

Developing Automated Algorithmic Options Trading Strategies advised by Prof. Nikunj Kapadia

February 2018 - Present

  • Developing modules that control decisions to enter a trade, quantity of trade, risk management and roll-over decisions.
  • Plan to optimize each module using machine learning and compare performance with a pre-decided benchmark.

Predicting Steering Angles in Self Driving Cars using Convolutional Neural Networks

October 2017 - December 2017

  • Designed and applied CNNs to predict the angles using images of the road captured from behind the car’s windshield.
  • Applied pre-trained VGG16 model and extended using Dropout and Dense layers to compare results of indigenous CNN.
  • Performed data augmentation using methods for shift, shadow and flip for a more generalized perception of the data.

Search Engine: Design and Implementation

October 2017 - December 2017

  • Implemented an end-to-end Information Retrieval Engine for structured query retrieval on Shakhesperean literature
  • Focused on performance evaluation, rank relevance order and efficient index creation for storage.

Assessing the impact of various factors on movie revenues with a focus on critic reviews

October 2017 - December 2017

  • Evaluated sentiments of movie reviews using Naïve Bayes and calculated sentiment scores using NLTK Vader.
  • Predicted revenues using Machine Learning models such as Decision Trees w/ AdaBoost, Logistic and Support Vector Regression on a dataset containing sentiment scores, budget, genres, ratings, release month, runtime etc. as features.

Distributed Home Automation System

February 2017 - May 2017

  • Simulated a distributed network of virtual devices to develop a smart home system.
  • Using RPCs for communication, the devices worked seamlessly with clock synchronization, fault tolerance, event ordering, consistency and consensus protocols.

Predicting Soccer League winners using Machine Learning

February 2017 - May 2017

  • Used feature selection and PCA to identify important features of self-curated dataset and their impact on a match’s score.
  • Incorporated ML models such as SVM, SGD etc. to predict results of individual matches and thus the winner of the league.
  • Studied the extent of effect of past seasons’ performance and team brand name on predictions of results.

Iterative and Feedback-based Data Cleaning for Statistical Modeling

February 2017 - May 2017

  • Processing and cleaning dirty datasets has always been a major bottleneck for any data analytics process. For a general Machine Learning prediction task, correctness in data implies higher final model accuracy.
  • Generally cleaning is done iteratively - clean some data, analyze and clean some more data based on the analysis. Unfortunately, for statistical models, iterative cleaning and re-training can lead to misleading results even for a simple model [Simpsons Paradox].
  • We explore an iterative data cleaning process in the context of training Machine Learning models that only enhances modeling accuracy at every step of the cleaning process by updating model parameters.
  • We further provide a novel way of integrating a feedback-mechanism from the Cleaner module and the updated ML model for choosing subsequent dirty samples for clean-up.

Internet of Things based Home Automation System using Raspberry Pi

September 2015 - May 2016

  • Designed a home automation system using Internet of Things with the central server(LAMP) as a Raspberry Pi.
  • Designed an indigenous module of wireless sensor networks that transmit the data collected to the Pi which integrates the data obtained from the sensors to automate the system.
  • Developed an Android app using JAVA to provide user with functionality to control various devices in the system.
  • The App and Raspberry Pi communicate with each other through a PHP framework.

Miscellaneous

Assistantships

  • Graduate Assistant, COMPSCI 520: Theory and Practice of Software Engineering - Graduate Level Course, UMass Amherst, Fall 2017
  • Graduate Assistant, COMPSCI 111: Introduction to Computer Science - Undergraduate Level Course, Smith college, Spring 2018

Extra-curricular Activities

  • Ed-Support Volunteer, Make a Difference: Taught English and Math to under privileged children in Nagpur, India.
  • Core-Committee member, E-Cell VNIT (2014-2015): In-charge of event planning, publicity and execution.
  • Selected as the Head of Corporate Relations, Consortium ’15 VNIT for sponsorships and events.