The project
B2H is building a Sharia-compliant trading assistant: a financial analysis platform that surfaces investment opportunities and explains every recommendation it makes.
The system combines technical analysis (support and resistance, Japanese candlesticks, indicators), machine learning models, risk management and Islamic finance rules. It produces BUY / HOLD / AVOID signals with an explanation attached — and never executes an order automatically. The decision stays human.
You join a team of three interns — AI, Machine Learning, Software — to build an MVP in one month.
Your mission
Build the machine learning models for financial market analysis and prediction.
Responsibilities
Data
- Collect and prepare historical market datasets
- Clean and preprocess financial data for training
- Build feature engineering pipelines from financial indicators
Financial features
- Detect support and resistance zones
- Analyse trends and volatility
- Extract price action features
- Identify Japanese candlestick patterns: hammer, doji, bullish engulfing, bearish engulfing, morning star, shooting star
Technical indicators
RSI · MACD · moving averages · Bollinger Bands · ATR · volume indicators
Modelling
Develop and compare several approaches: logistic regression, Random Forest, XGBoost, LightGBM, and LSTM as an option.
Evaluation
- Accuracy, precision, recall, F1-score, confusion matrix
- Backtesting, win rate, risk/reward ratio, maximum drawdown
Deliverables
- Financial data pipeline
- Feature engineering module
- Machine learning models
- Model evaluation report
- Backtesting results
Who we are looking for
- Studying data science, statistics, computer science or equivalent
- Python and its data ecosystem (pandas, scikit-learn)
- An understanding of evaluation metrics and the traps of overfitting
- Methodological honesty: on financial time series, a data leak produces flattering scores and a useless model
- Familiarity with financial markets is a plus, not a prerequisite
Timeline
A one-month internship in four stages: framing and research, core development, integration with the other components, then testing, evaluation and the final MVP presentation.
How to apply
Fill in the form below with your CV. We reply within 48 business hours.