AI Based Table Tennis Predictive Model

I'm interested in developing a predictive model for table tennis games. This model should be designed using state-of-the-art AI algorithms, and be fuelled by live scraped data. Further details include:

* Data for Model: Primarily, live scraping data will be utilized to train the predictive model. No historical data is required, meaning that the model must be capable to learn and adapt predominantly from ongoing games.

* Update Frequency: It's crucial for the system to update the data used for predictions constantly. The predictive model ought to be dynamic and should be able to continuously learn and predict from live results.

* Key Features: The predictor must focus on match results fetched from the live scraping. No additional variables such as player rankings or match locations is needed.

In terms of skills required for this project, the freelancer needs to have a solid background in AI, specifically in predictive modeling. They should also have experience in live data scraping, and ideally have prior knowledge of table tennis to understand the key elements of the match outcomes. Working with dynamic models and real-time data would be beneficial.