Pharma Demand Forecasting Tool

I need an expert to help enhance our Demand Forecasting Tool for the pharmaceutical business. Our current plan involves:

- Initial forecasts using Time Series models (ARIMA, Prophet, ETS)
- Using these forecasts as a baseline for a Reinforcement Learning (RL) model
- Making forecasts more accurate day by day
- Developing a continuous learning RL model

My current SL and RL models are ready but struggle with high variability in new datasets.

Key tasks include:
- Adjusting parameters and the reward function to work well across all datasets
- Ensuring the model can adapt and learn continuously

Ideal skills and experience:
- Strong background in time series forecasting and reinforcement learning
- Experience with high variability datasets
- Proficiency in adjusting model parameters and reward functions
- Ability to develop robust continuous learning systems

Focus on Mean Absolute Error (MAE) for performance metrics.