About the job
Aspire Software is hiring a Data Scientist with a focus on reinforcement learning for its team in Beirut, Lebanon.
About Aspire Software
Aspire Software manages a group of wholly owned software companies that provide mission-critical solutions across several industries. The company emphasizes industry best practices and an efficient integration process. Its decentralized structure supports rapid growth and ongoing reinvestment in a diverse portfolio.
Role Overview
This role centers on moving machine learning models, especially reinforcement learning, from research to deployment. Aspire Software’s current pricing engine relies on rule-based parametric systems, using elasticity modeling, sigmoid demand curves, and occupancy deviations. The next step is to transform this engine into a dynamic learning system. The work involves contextual bandits, reinforcement learning policies, and probabilistic forecasting to set prices for thousands of hotel-room nights each day. External factors such as competitor pricing, local events, and weather will also feed into the pricing strategy.
What You Will Do
- Pricing Intelligence: Upgrade the parametric pricing engine using advanced models, including contextual bandits and Bayesian elasticity estimation.
- Reinforcement Learning in Production: Design reward functions, exploration strategies, and off-policy evaluation methods to deploy reinforcement learning safely across hotel data from multiple tenants.
- Demand Forecasting: Improve booking-curve and occupancy forecasts with advanced time-series and probabilistic techniques.
- Simulation & Evaluation: Build out the historical replay and synthetic simulation framework to support offline evaluation and A/B testing of pricing policies.
- LLM-Powered Features: Develop intelligent workflows that use advanced AI models for event-based pricing recommendations and demand analysis.
- Productionization: Write production-ready Python services using FastAPI and SQLAlchemy, focusing on quality and maintainability.
