Viral Pandey

Cambridge, MA · pandey.v@northeastern.edu

Data Scientist at nference. Previous Data Scientist at Tesla.

Experience

Data Scientist Intern

Tesla, Inc.
  • Developed supervised regression models to predict congestion and determine the capacity expansion of Supercharger sites
  • Built data pipelines to convert vector data of public roads into Uber’s H3 hexagons. This helped me design and put Traffic Coverage and Road Coverage KPIs into production
  • Designed a time series forecasting model to estimate quarterly energy usage at sites. This informed the estimation of $ revenue from the entire Supercharger network for future quarters
  • Quantified the population coverage of the world using geo-spatial data of population density per pixel of the world and isochrone coverage (areas within some minutes by driving) of sites
  • Identified vehicles that might be involved in potential misuse of the Supercharger network. Proposed false positive scenarios as well as solutions to mitigate such incidents

Aug - Dec 2019 | May - Aug 2020

Research Assistant

Khoury College, Northeastern University
  • Explored Procedure Learning problem of understanding the constituting key actions of complex tasks from instructional video data
  • Assembled a Fully Convolutional Sequential Network (FCSN) that produces a compact summary of the procedure steps and their ordering needed to perform a complex task, as well as localization of these steps in videos
Jul 2019

Teaching Assistant

Khoury College, Northeastern University
  • Worked closely with the professors to reinforce class material by answering questions about concepts and assignments from students and help them practice effective problem-solving techniques
  • Held weekly office hours, graded programming assignments and exams. I have been the TA for the following courses:
    • CS6140 Machine Learning
    • DS5110 Data Management and Processing
    • DS2000 Programming with Data
Jan - Jun 2019 | Jan - Apr 2020

Data Science Research

Dhirubhai Ambani Institute of Information and Communication Technology
  • Outperformed other algorithms in forecasting Remaining Useful Life of a jet engine based on NASA′s time series dataset by developing a Recurrent Convolutional Neural Network (RCNN) based predictive model
Jan - Apr 2018

Data Engineer Intern

Infoware
  • Implemented Hadoop infrastructure and used Hive on historical data of client′s customers across different sales channels
May - July 2017

Education

Northeastern University

Master of Science in Data Science
GPA: 3.89/4.00
Relevant Coursework:
  • Causal Inference
  • Deep Learning
  • Supervised Machine Learning
  • Unsupervised Machine Learning
  • Foundations of Artificial Intellligence
  • Algorithms
  • Data Management and Processing
Aug 2018 - Dec 2020

Dhirubhai Ambani Institute of Information and Communication Technology

Bachelor of Technology in Information and Communication Technology
GPA: 7.04/10.00
Relevant Coursework:
  • Selected Topics in Neural Networks
  • Database Management System
  • Stochastic Simulation, Models of Computation
  • Object Oriented Programming
  • Data Structures
Aug 2014 - May 2018

Projects

Named Entity Recognition (NER) and Relation Extraction (RE) from Patient’s Medical Notes

  • Highlighted entities like Drugs, Adverse effect, Dosage, Reason, etc and mapped the Drug entity with all other entities to create a structured data table out of unstructured notes
  • Achieved 90% micro-F1 score for NER and RE using BioBERT and BiLSTM+CRF models
  • Built a website and APIs to get model predictions using FastAPI and hosted them on Google Cloud Platform
Jan - Mar 2019

Quora Insincere Question Classification

  • Designed a supervised binary classifier to detect insincere questions on the Quora website and compared performances of algorithms such as SVM, CNN and LSTM RNN
  • Performed TF-IDF vectorization, Sentiment Analysis using Python NLTK framework for gauging overall sentiment
Jan - Mar 2019

Bankruptcy Prediction Using Various Classifiers

  • Tested Logistic Regression, Naive Bayes, LDA, QDA, SVM and feed forward Neural Networks on 5 years of Polish Companies Data of different econometric ratios to determine whether a company is going to go bankrupt
  • Handled missing data using Mean value Imputation and SMOTE on training data to handle class imbalance
  • Calculated Correlation matrix, performed PCA and cross validation for feature sub-setting
Jan - Mar 2019

Analysis of NHIS data and building associated software components

  • Developed a R package, performed exploratory data analysis to find correlation between different variables
  • Developed a conversational chat bot which can show the analysis as visualizations using Node.js, R, Dialogflow and AmazonS3Services
Nov - Dec 2018

Skills

Programming Languages
Tools & Libraries
  • TensorFlow, Keras, PyTorch
  • Pyro
  • OpenCV
  • MySQL
  • Postgres
  • Google Cloud Platform
  • Hive
  • RStudio
  • Jupyter
  • Google Colab
  • Git
Workflow
  • Cross Functional Teams
  • Agile Development & Scrum

Interests

Outside of my class, I try to attend hackathons and workshops happening on or around my university campus. I have participated in multiple hackathons like IBM - State Street Hackathon and AT&T Entertainment Hackathon to develop apps using technologies like IBM Watson Vision Recognition, Node-RED, Azure Face API, Heroku and Spotify API.

At DAIICT (my undergrad university), I was a member of the Google Developers Group, DAIICT Chapter. We organized different tech events where we schedule lectures, hands on events, tutorials and live broadcast events. I was also an Organizing committee member of YouthRun, our Annual techno-cultural fest. As a coordinator, I handled large groups of volunteers, conducted meetings with other members, approached various companies for sponsorships and managed various production side objectives.

In my pastime, I enjoy being outdoors. I like taking frequent road trips to near by places and during summer months, I enjoy biking and kayaking. When forced indoors, I follow a number of sci-fi and fantasy genre movies and television shows.

Contact

I'm glad you made it this far! Reach out to me via email, text or tweet.