AI-Powered Quant Risk Desk

A full-stack quantitative risk management platform that combines derivative pricing, portfolio analytics, market risk measurement, stochastic volatility modeling, and AI-generated risk reporting.

This project simulates the workflow of a modern derivatives risk desk by integrating yield curve construction, volatility estimation, option pricing engines, portfolio valuation, risk analytics, stress testing, Value-at-Risk (VaR), Expected Shortfall (CVaR), and LLM-powered risk commentary.

---

Project Overview

Modern quantitative risk desks perform much more than option pricing. They continuously:

• Build and maintain market curves
• Estimate volatility and risk factors
• Price derivative instruments
• Compute risk sensitivities (Greeks)
• Perform stress testing
• Measure Value-at-Risk (VaR)
• Monitor portfolio risk
• Generate risk reports for management

This project replicates that workflow in a simplified but realistic environment.

---

Key Features

Market Data Engine

• Excel-based market data ingestion
• Yield curve construction
• Discount factor calculation
• Forward rate generation
• Historical volatility estimation

Pricing Engines

Black-Scholes (C++)

• European call and put pricing
• Fast C++ implementation
• Python integration wrapper

Monte Carlo Pricing

• Risk-neutral simulation
• Option valuation using simulated paths
• Benchmark comparison against Black-Scholes

Heston Stochastic Volatility Model

• Stochastic volatility path generation
• Monte Carlo option pricing under Heston dynamics
• Market-volatility calibration using historical volatility estimates

Portfolio Analytics

• Multi-position option portfolio valuation
• Portfolio aggregation
• Position-level valuation reporting

Greeks Engine

Portfolio-level:

• Delta
• Gamma
• Vega
• Theta
• Rho

Risk Analytics

Stress Testing

Scenario analysis including:

• Spot price shocks
• Volatility shocks
• Interest rate shocks

Value-at-Risk (VaR)

Implemented using:

• Historical Simulation
• Parametric VaR
• Monte Carlo VaR

Expected Shortfall (CVaR)

Implemented using:

• Historical ES
• Monte Carlo ES

Volatility Surface

• Synthetic implied volatility surface generation
• Interactive 3D visualization using Plotly

AI Risk Analyst

Integrated with:

• Ollama
• Llama 3.1

Features:

• Portfolio risk assessment
• Tail risk commentary
• Pricing model analysis
• Risk management recommendations

Automatic fallback to a rule-based analyst when Ollama is unavailable.

Reporting

• Automated PDF report generation
• AI-generated markdown reports
• Interactive Streamlit dashboard

---

Technology Stack

Quantitative Finance

• Yield Curve Construction
• Discount Factors
• Forward Rates
• Black-Scholes Model
• Monte Carlo Methods
• Heston Model
• Greeks
• Value-at-Risk
• Expected Shortfall
• Stress Testing

Programming

• Python
• C++

Libraries

• Pandas
• NumPy
• Plotly
• Streamlit
• ReportLab
• Requests

AI

• Ollama
• Llama 3.1

---

Project Structure

Quant-Risk-Desk/

├── ai_engine/
├── curve_engine/
├── dashboard/
├── data/
│   ├── raw/
│   └── processed/
├── heston_engine/
├── pdf_reporting/
├── portfolio_engine/
├── pricing_engine_cpp/
├── pricing_wrapper/
├── risk_engine/
├── volatility_engine/

├── main.py
├── requirements.txt
└── README.md

---

Dashboard

The dashboard provides:

• Portfolio valuation
• Greeks visualization
• Stress testing analysis
• VaR and Expected Shortfall metrics
• Volatility surface visualization
• Pricing model comparisons
• AI-generated risk memo
• PDF report download

---

Running the Project

Clone Repository

git clone <repository-url>
cd Quant-Risk-Desk

Create Virtual Environment

python -m venv venv

Activate

Windows:

venv\Scripts\activate

Install Dependencies

pip install -r requirements.txt

Run Analytics Pipeline

python main.py

Launch Dashboard

streamlit run dashboard/app.py

---

Optional AI Setup

Install Ollama:

https://ollama.com

Download model:

ollama pull llama3.1

Verify:

ollama run llama3.1

The project automatically falls back to the built-in rule-based analyst if Ollama is unavailable.

---

Future Improvements

Potential extensions:

• SABR Model
• Local Volatility Model
• Hull-White Interest Rate Model
• LIBOR Market Model
• Portfolio Optimization
• Delta Hedging Simulator
• Real-Time Market Data Integration

---

Author

Built as a quantitative finance, risk management, and AI engineering project to demonstrate skills relevant to:

• Quantitative Research
• Quantitative Trading
• Quantitative Development
• Risk Management
• Financial Engineering
• AI for Finance