Vishal Morde - Databricks

Vishal Morde

Data Scientist, Uber, Inc. 

An award-winning Data Science leader with 16+ years of leadership experience in global technology and financial institutions. Equal parts visionary, strategist, and hands-on tactician with demonstrated success in leading enterprise-wide data/analytics transformation. A distinguished thought leader, keynote speaker, and influencer in Artificial Intelligence, Machine Learning, Data Strategy, & Advanced Analytics.


How Much are Your Customers Really Worth? Measuring Customer life-time Value using Deep LearningSummit 2020

At Uber, we ignite opportunity by setting the world in motion. 100MM+ customers across 60+ countries use our Rides, Eats, Freight, Bikes, Scooter services and we are investing heavily in new modes of transportation including Uber Air and autonomous vehicles. As our business matures, we have been continuously refining the customer lifetime value (LTV) framework to improve the financial discipline and capital allocation strategies. Our research in AI and Machine Learning has allowed us to leverage Deep Learning, specifically Recurrent Neural Networks, to build a highly accurate and scalable valuation framework. The presentation will provide an overview of the model architecture, data processing, feature engineering, algorithm development, and model deployment process. We will share the results from the existing model in production that doubled accuracy while maintaining scalability and interpretability. Key Takeaways:

  • Lessons learned from building state-of-the-art customer life-time valuation models
  • Current trends in Deep Learning and use cases in financial forecasting, business strategies, and strategic management
  • Best practices in using Apache Spark for analyzing billions of data points from every stage of the customer lifecycle
  • Future trends in enterprise-wide AI-powered decision making