AI Market Trend Predictor

I need an AI model that can look at current market prices in the Tree Market and turn that real-time feed into daily forecasts I can rely on. The core job is to build, train and deploy a prediction engine—nothing else gets the same priority.

Here is what I have in mind:

• Data scope: the system should ingest only today’s live price data (no historical or social signals for now) and transform it into usable features automatically.
• Update cadence: every 24 hours the model must refresh its dataset, retrain or recalibrate if needed, and publish a new set of market-trend predictions.
• Output format: a concise JSON or CSV report with price direction, confidence score and any key indicators you derive.

I am comfortable with Python and would prefer to host the solution on a lightweight cloud instance (AWS or similar), but suggest alternatives if they shorten turnaround time. Please include a short note on the framework you plan to use—TensorFlow, PyTorch, or a lighter ML library are all fine as long as they support scheduled retraining.

Acceptance criteria
1. An executable script or container that fetches current price data, processes it, and stores the fresh prediction daily.
2. A README that explains setup, scheduling and any environment variables.
3. A quick demo run showing one full cycle from data pull to prediction generation.

Let me know your approach and the estimated timeframe to get the first working model live.