Quant AI Research Platform

A production-grade quantitative research platform for systematic investing, feature engineering, machine learning, portfolio optimization, and backtesting on Indian financial markets.

The platform automates the complete quantitative research lifecycle:

• Market data ingestion
• Data validation
• Feature engineering
• Alpha signal generation
• Machine learning modeling
• Portfolio construction
• Backtesting
• Performance analytics
• Experiment tracking

Built using modern Python engineering practices and institutional-grade tooling.

---

Features

Market Data Ingestion

Supported Data Sources:

• YFinance
• Alpha Vantage
• FRED
• News API

Capabilities:

• Historical data collection
• Incremental updates
• Retry logic
• Rate limiting
• Data validation
• Schema enforcement
• Automatic failure recovery

---

Feature Engineering

Trend Features

• SMA 5
• SMA 10
• SMA 20
• SMA 50
• SMA 200
• EMA 10
• EMA 20
• EMA 50

Momentum Features

• RSI 14
• MACD
• MACD Signal
• ROC 5
• ROC 10
• Momentum 5
• Momentum 10

Volatility Features

• Volatility 10
• Volatility 20
• Volatility 50
• ATR 14
• Bollinger Band Width

Mean Reversion Features

• Z Score 20
• Distance from SMA
• Distance from EMA

Volume Features

• Volume MA 20
• Volume Ratio
• On Balance Volume (OBV)

Targets

• Future Return 1 Day
• Future Return 5 Days
• Future Return 10 Days

---

Architecture

                ┌──────────────────┐
                │ External Sources │
                └─────────┬────────┘
                          │
        ┌─────────────────┼─────────────────┐
        │                 │                 │
        ▼                 ▼                 ▼

   YFinance      AlphaVantage        FRED API
        │                 │                 │
        └─────────────────┴─────────────────┘
                          │

                Data Connectors Layer
                          │
                          ▼

                Data Validation Layer
                          │
                          ▼

                 Raw Market Dataset
                          │
                          ▼

                Feature Engineering
                          │
                          ▼

                 Feature Dataset
                          │
                          ▼

                 Target Generation
                          │
                          ▼

                 ML Training Layer
                          │
                          ▼

                 Signal Generation
                          │
                          ▼

              Portfolio Construction
                          │
                          ▼

                   Backtesting
                          │
                          ▼

               Performance Metrics
                          │
                          ▼

                      MLflow

---

Project Structure

quant_ai_platform/

├── config/
│   ├── constants.py
│   ├── settings.py
│   └── __init__.py
│
├── data/
│   ├── raw/
│   ├── processed/
│   ├── features/
│   ├── predictions/
│   └── signals/
│
├── alembic/
│
├── docker/
│
├── notebooks/
│
├── src/
│
│   ├── data_pipeline/
│   │   ├── connectors/
│   │   │   ├── yfinance_connector.py
│   │   │   ├── alpha_vantage_connector.py
│   │   │   └── fred_connector.py
│   │   │
│   │   ├── validation/
│   │   └── processing/
│   │
│   ├── feature_engineering/
│   │   ├── trend_features.py
│   │   ├── momentum_features.py
│   │   ├── volatility_features.py
│   │   ├── mean_reversion_features.py
│   │   ├── volume_features.py
│   │   ├── target_generator.py
│   │   └── feature_pipeline.py
│   │
│   ├── models/
│   │   ├── asset.py
│   │   ├── price.py
│   │   ├── feature.py
│   │   ├── signal.py
│   │   ├── trade.py
│   │   └── portfolio_metric.py
│   │
│   ├── database/
│   │   ├── base.py
│   │   ├── session.py
│   │   │
│   │   └── repository/
│   │       ├── asset_repository.py
│   │       ├── price_repository.py
│   │       ├── feature_repository.py
│   │       ├── signal_repository.py
│   │       ├── trade_repository.py
│   │       ├── portfolio_metric_repository.py
│   │       └── __init__.py
│   │
│   ├── utils/
│   │   └── logger.py
│   │
│   ├── training/
│   ├── backtesting/
│   ├── portfolio/
│   └── api/
│
├── tests/
│   ├── database/
│   ├── feature_engineering/
│   ├── data_pipeline/
│   └── api/
│
├── docker-compose.yml
├── pyproject.toml
├── .env.example
└── README.md

---

Database Schema

Assets

Stores security master data.

Column	Description
symbol	Trading symbol
name	Asset name
asset_class	Equity, ETF, Index, etc
exchange	NSE, BINANCE, FX
currency	Trading currency
sector	Industry sector
is_active	Active flag

---

Prices

Stores OHLCV market data.

Column	Description
symbol	Asset symbol
date	Trading date
open	Open price
high	High price
low	Low price
close	Close price
adj_close	Adjusted close
volume	Trading volume
source	Data source

---

Features

Feature metadata registry.

Column	Description
symbol	Asset symbol
date	Feature date
feature_set_version	Feature version

---

Signals

Model predictions.

Column	Description
symbol	Asset symbol
date	Signal date
signal	Buy/Hold/Sell
predicted_return	Expected return
model_version	Model version

---

Trades

Backtest execution records.

Column	Description
symbol	Asset symbol
entry_date	Entry timestamp
exit_date	Exit timestamp
entry_price	Entry price
exit_price	Exit price
pnl	Profit/Loss
return_pct	Percentage return

---

Portfolio Metrics

Strategy performance tracking.

Column	Description
strategy_name	Strategy identifier
metric	Metric name
value	Metric value
calculated_at	Timestamp

---

Setup

Clone Repository

git clone https://github.com/<username>/quant_ai_platform.git

cd quant_ai_platform

---

Create Virtual Environment

Linux / macOS

python3.12 -m venv .venv

source .venv/bin/activate

Windows

python -m venv .venv

.venv\Scripts\activate

---

Install Dependencies

pip install -U pip

pip install -e .

---

Configure Environment Variables

Copy:

cp .env.example .env

Update values:

DATABASE_URL=postgresql://postgres:postgres@localhost:5432/quantdb

ALPHA_VANTAGE_KEY=
FRED_API_KEY=
NEWS_API_KEY=

DATA_START_DATE=2015-01-01
DATA_UPDATE_MODE=incremental

MARKET=NSE
STOCK_UNIVERSE=NIFTY500

---

Infrastructure

Start PostgreSQL and MLflow:

docker compose up -d

Verify:

docker ps

---

Database Migrations

Create migration:

alembic revision --autogenerate -m "initial schema"

Apply migration:

alembic upgrade head

Rollback:

alembic downgrade -1

---

Running Data Pipeline

Run data ingestion:

python -m src.data_pipeline.connectors.yfinance_connector

Run feature generation:

python -m src.feature_engineering.feature_pipeline

---

Running Tests

Run all tests:

pytest

Run database tests only:

pytest tests/database -v

Run with coverage:

pytest --cov=src --cov-report=term-missing

---

MLflow

Launch:

docker compose up -d mlflow

Open:

http://localhost:5000

Track:

• Experiments
• Parameters
• Metrics
• Models
• Artifacts
• Feature versions

---

Development Roadmap

Phase 1 — Research Infrastructure

• Database Models
• Repository Layer
• Data Connectors
• Feature Engineering Pipeline
• Automated Testing
• MLflow Integration

Phase 2 — Machine Learning

• Feature Store
• XGBoost Models
• LightGBM Models
• Model Registry
• Hyperparameter Optimization

Phase 3 — Signal Engine

• Signal Generation
• Ensemble Models
• Model Monitoring

Phase 4 — Portfolio Construction

• Position Sizing
• Risk Controls
• Portfolio Optimization
• Exposure Constraints

Phase 5 — Backtesting

• Transaction Costs
• Slippage Modeling
• Walk Forward Testing
• Performance Analytics

Phase 6 — Production

• FastAPI Service
• Dashboard
• Scheduled Pipelines
• Paper Trading
• Live Deployment

---

Engineering Principles

• SQLAlchemy 2.0 Typed ORM
• Repository Pattern
• Dependency Injection
• Alembic Migrations
• Transactional Test Isolation
• Type Safety
• Production Logging
• Dockerized Infrastructure
• Reproducible ML Experiments
• No Lookahead Bias
• Institutional Research Standards

---

Disclaimer

This project is intended for research and educational purposes only.

Nothing contained in this repository constitutes investment advice, financial advice, trading advice, or a recommendation to buy or sell securities.

All investments involve risk, including the possible loss of principal.

---

License

MIT License

Copyright (c) 2026