Skip to content
Upasana Raval
Upasana Raval

Upasana Raval

Operations Research practitioner
American Airlines
Bio

Upasana Raval is an Operations Research practitioner, holding dual master’s degrees from the University of Wisconsin-Madison. She has developed advanced analytics tools for companies like American Airlines, Memorial Sloan Kettering Cancer Center, Nestle, and Schneider National, by leveraging her specialization in optimization modeling and predictive analytics. Currently, at American Airlines, she builds tools that improve customer experience, optimize flight schedules, predict revenue, and allocate maintenance resource optimally.


An Ensemble Machine Learning Engine for Predicting Flight Revenue to Enable Dynamic Flight Scheduling at American Airlines: From Concept to Deployment

Featuring: Yanqing Kuang

At American Airlines, we’re transforming how flight schedules are planned and evaluated through a cutting-edge platform developed in collaboration with Palantir. This “one-stop shop” empowers planners and schedulers to assess flight schedules in minutes—balancing operational feasibility, profitability, and robustness.

A key component of this platform is the engine that predicts flight-level revenue, enabling analysts from crew and network planning teams to assess the profitability of flight schedules dynamically with respect to changes. The engine uniquely leverages concepts from optimization, econometrics, and machine learning to provide flight revenue prediction with 95% accuracy. In this session, we’ll walk through our end-to-end engine development and deployment journey—from training ensemble CatBoost models and interpreting results using SHAP, to leveraging Palantir Foundry for scalable experimentation and deployment, and packaging the core logic for seamless integration across platforms and applications

Essential / Professional