Predictive Trading Mobile App
Budget: $250 – $750 USD
I need a mobile app built for both iOS and Android that helps me decide whether to place a call or put order on a chosen stock or on Bitcoin. The core of the app is a machine-learning model that must output direction predictions over three short-term horizons—1, 2, and 5 minutes—and refresh fast enough for real-time use.
Key trading logic
– Order size presets of 1 %, 2 %, 3 %, or 5 % of the current account balance, with the option to override the preset by entering an exact dollar amount.
– Live price feeds and trade execution against NASDAQ, NYSE, and Coinbase, so the same interface works whether I pick a listed stock or Bitcoin.
– A running trade ledger that stores date/time, symbol, horizon chosen, call/put side, stake size, and the eventual P&L, then displays summary statistics I can filter later.
Please implement the ML pipeline on the back end (you are free to suggest the specific libraries, but TensorFlow Lite or Core ML compatibility will help on-device inference). The model does not have to be perfect out of the gate; a basic, reproducible framework with clear hooks for retraining is more important.
Deliverables
• Universal iOS + Android app (single code-base preferred)
• Deployed ML model producing 1-, 2-, and 5-minute predictions
• Secure brokerage/exchange integration for NASDAQ, NYSE, and Coinbase
• Trade history module with export to CSV/Excel
• Source code, build instructions, and short user guide
Acceptance criteria will be a working demo that connects to live market data, allows test trades in a sandbox environment, and logs at least ten sample transactions end-to-end.
Key trading logic
– Order size presets of 1 %, 2 %, 3 %, or 5 % of the current account balance, with the option to override the preset by entering an exact dollar amount.
– Live price feeds and trade execution against NASDAQ, NYSE, and Coinbase, so the same interface works whether I pick a listed stock or Bitcoin.
– A running trade ledger that stores date/time, symbol, horizon chosen, call/put side, stake size, and the eventual P&L, then displays summary statistics I can filter later.
Please implement the ML pipeline on the back end (you are free to suggest the specific libraries, but TensorFlow Lite or Core ML compatibility will help on-device inference). The model does not have to be perfect out of the gate; a basic, reproducible framework with clear hooks for retraining is more important.
Deliverables
• Universal iOS + Android app (single code-base preferred)
• Deployed ML model producing 1-, 2-, and 5-minute predictions
• Secure brokerage/exchange integration for NASDAQ, NYSE, and Coinbase
• Trade history module with export to CSV/Excel
• Source code, build instructions, and short user guide
Acceptance criteria will be a working demo that connects to live market data, allows test trades in a sandbox environment, and logs at least ten sample transactions end-to-end.