Marketing Science Analyst, Annalect, New York April 2019 - Present • Built an interactive R shiny web application for a client that would allow them allocate optimized budgets to different marketing channels • Led a team to automate the entire framework that would perform data manipulations, data preparations, modeling and make automated updates on the product improving run time and availability by 150 percent • Designed an optimized Markov chain algorithm to analyze market performance increasing the speed by 5 times • Designed an algorithm that would provide optimized model results addressing the data vulnerability • Designed interactive sanky charts and heat maps using D3.js to obtain better insights into user conversion behavior • Performed data modeling and custom product design for winning client pithes resulting in huge revenue increase • Manage day to day product improvements and feature updates catering to stakeholder requirements • Built custom R packages that can be used to run processes in a quick and efficient manner
Cloud Information Management and Machine Learning Intern, Axtria, New Jersey June 2018 - Dec 2018 • Led a team to design and develop a UI with tableau and SQL using real world insurance claims data, allowing users to apply pharma-specific filters and obtain final cohort for analysis, resulting in $1m in revenue every year • Analyzed data to provide statistical visualizations on final cohort obtained after applying filters on the tool • Designed data ingestion and transformation pipelines using AWS that eased data availability for the developed UI • Analyzed real world marketing and sales data and designed a machine learning model using random forest and neural network that would suggest optimal future promotions to be made by pharmaceutical companies • Designed a machine learning web visualization for the above analysis using D3.js, providing deeper insights and intuition into the problem and solution with interactive visuals and graphs • Designed dashboards to analyze AWS service’s and spark logs using Elasticsearch-Logstash-Kibana stack that improved system governance and quality check for client infrastructure • Designed AI/ML use cases and solutions to optimize pharma company’s marketing and sales processes • Analyzed internal company data to provide deeper insights into company’s progress over time • Led a team in developing proposal presentation that were later presented to the client
Projects
Instacart Market Basket Analysis. Apr 2018 – May 2018
Project descriptionInstacart is an American online grocery delivery service 1. Implemented plethora of feature engineering on Instacart user data to create 23 new features that would give optimal results while modeling 2. Executed modeling using Xgboost to predict what product will a user order next to implement auto-carting feature on the website 3. Performed market basket analysis to provide appropriate user suggestions while they shop to enhance customer experience
Mode-Separation Generative Adversarial Networks Jan 2018 – May 2018 1. Explored the mode-separation problem in the Gans framework and premeditated various solutions to deal with it 2. Devised an approach that used multiple generators in the network each learning a different distribution from the input data 3. Deploying a pre-trained classifier with the discriminator forced each generator to learn a different distribution
Airbnb New User Destination Prediction Mar 2018 – Apr 2018 1. Cleaned, merged and normalized the Airbnb dataset consisting of 3 tables to make it optimal for modeling 2. Built a model to predict the next user destination by using an amalgamation neural networks and xgboost algorithms 3. Analyzed the model and provided insights and suggestions to the marketing team for further use
Persona Modeling Jan 2018 – Feb 2018 1. Merged different datasets in R to create a complete university dataset for analysis 2. Cleaned and normalized the dataset to make it optimal for further computations 3. Imposed data augmentation and prediction modeling techniques to fill in the missing values in the dataset 4. Built various disjoint personas using R and Tableau that can be used further by the university Marketing team for attracting new students
Models in Machine Learning and Data Analytics Sep 2017 – Dec 2017 1. Developed and analyzed a linear regression model to predict the power consumption by predicting the current drawn into the joints as a function of the robot motion, achieving an efficiency of 85 per cent 3. Implemented a multi-class classification model to analyze a breast cancer dataset 4. Developed a multi-class classification model to predict a few characteristics in mice by measuring expression levels across 77 genes achieving 97 per cent accuracy 5. Developed and Analyzed a model using support vector machine to identify a particular digit or alphabet 6. Developed an optimization model to find the pitch in an audio signal