Keshav Ramaswamy

Infor Cambridge,MA

Data Scientist • Sep, 2016 — Current

I am a part of the Dynamic Science Labs team where we have worked on building science features for enterprise software and implementing customized solutions for our customers.

• Infor CRM Sales Intelligence
  • Designed and developed models for New Prospect Scoring, Cross Sell Scoring and Recommender System based on Transactions and Engagement Data using various Machine Learning techniques.
  • Product Adoption resulted in 2x increase in New Prospect conversion performance with more than 70% of the conversions from 20% of the leads.
  • Cross Sell Recommendations also enabled diversification of customer's business and ~6% increase in sales revenue from less popular products.
  • Led the design, development and implementation of the data pipeline which powers the RShiny Dashboard for the clients.
  • Primary collaborator with the Engineering team migrating to production.
• Internal Customer Churn Scoring
  • Led the design and development of a ML based internal solution to identify customers at high risk of churning and provide customer specific reasons impacting the risk. Collaborated closely with the Subscription Services team to work on improving the model predictions on an iterative basis.<\li>
  • Model has been able to capture 75% customer churn, resulting in 5x uplift over existing manual churn-risk evaluation, during back-testing and identify individualized business metrics impacting churn.
  • Currently working on implementing a closed loop process to alert Subscription Services Managers of high risk churning customers.
• Distributor Van Timeliness Prediction
  • Designed and developed a model which can identify delivery vans at a high risk of arriving late.
  • Historical Sales, GPS Data, Van Routes and schedules were used for modelling and analyze main factors causing late van deliveries.
  • 89% Late Deliveries were predicted to be late during back-testing - resulting in a very positive feedback from the client.