Quant Developer – Volatility Forecasting & Market Regime Detection Framework

Project Description Quant Developer – Volatility Forecasting & Market Regime Detection Framework Project Type: Quantitative Research & Development Domain: Systematic Trading | Multi-Strategy Portfolio Construction Project Background We are building an institutional-grade systematic trading platform comprising multiple alpha sleeves, primarily Trend Following and Mean Reversion, across liquid global futures markets. While each sleeve generates independent alpha under different market conditions, our research has identified a structural portfolio risk: During strong directional markets, trend-following and mean-reversion strategies often become highly correlated, leading to concentration of risk rather than diversification. To address this, we are seeking an experienced Quantitative Developer / Quant Researcher to design and implement a reusable Portfolio Intelligence Layer consisting of: • A Volatility Forecasting Engine for dynamic position sizing. • A Market Regime Detection Engine for dynamic allocation across strategy sleeves. This project is not focused on generating new alpha. Instead, it aims to improve portfolio construction, capital allocation, and overall risk-adjusted returns by making portfolio weights adaptive to prevailing market conditions. Project Objectives The successful candidate will design and validate a framework capable of: 1. Forecasting future market volatility for systematic position sizing. 2. Detecting changing market regimes in real time. 3. Dynamically allocating capital between Trend Following and Mean Reversion strategies based on identified regimes. 4. Reducing portfolio risk arising from sleeve correlation during persistent trends. 5. Building modular components that can be extended with future models and methodologies. Scope of Work 1. Volatility Forecasting Engine Develop a robust volatility forecasting framework that will be used as the portfolio's position sizing engine. The objective is to generate stable, forward-looking volatility estimates that can scale exposure across instruments while maintaining consistent portfolio risk. Potential methodologies include (recommendations welcome): • Rolling Realized Volatility • Exponentially Weighted Moving Average (EWMA / RiskMetrics) • GARCH family models o GARCH o EGARCH o GJR-GARCH • Other statistically robust forecasting techniques Responsibilities • Evaluate multiple forecasting approaches. • Compare models using: o Forecast accuracy o Stability o Responsiveness o Out-of-sample performance • Recommend the most appropriate methodology. • Produce a reusable volatility forecast series for every instrument. • Document assumptions, parameters, and implementation choices. • Integrate the output into a target-volatility position sizing framework. Expected Output • Modular volatility forecasting library • Forecast series for each instrument • Position sizing methodology • Documentation and implementation guide 2. Market Regime Detection Engine Design and implement a framework capable of classifying markets into distinct behavioral regimes. The primary objective is to dynamically adjust allocations between Trend Following and Mean Reversion based on prevailing market conditions. Minimum regime classifications should include: • Trending • Mean-Reverting • Choppy / Low Signal Methodology is open and may include: • Hidden Markov Models (HMM) • Rule-based classifiers • Trend strength indicators • Autocorrelation measures • Volatility clustering • Machine Learning classifiers • Hybrid approaches Key Requirement The framework must explicitly identify periods where Trend Following and Mean Reversion become highly correlated and systematically reduce portfolio concentration during such periods. Instead of maintaining fixed sleeve allocations (e.g., 50/50), the model should generate adaptive weights based on current market conditions. Responsibilities • Develop the regime classification framework. • Build a dynamic sleeve allocation model. • Validate regime transitions. • Measure regime persistence. • Quantify impact on sleeve correlation. • Provide confidence measures where appropriate. Expected Output • Regime classification engine • Dynamic allocation model • Historical regime labels • Allocation recommendations over time 3. Portfolio Integration & Validation Integrate both modules into the existing systematic trading framework. The integrated system should be benchmarked against the current static allocation methodology. Performance analysis should include: Portfolio Performance • CAGR • Annualized Return • Sharpe Ratio • Sortino Ratio • Calmar Ratio • Maximum Drawdown • Volatility • Return Distribution Portfolio Construction • Sleeve allocation history • Rolling sleeve correlation • Exposure changes • Turnover introduced by dynamic weighting Robustness • Parameter sensitivity analysis • Out-of-sample validation • Stability across different market environments • Discussion of assumptions and limitations The codebase should remain modular so that volatility models or regime detection methods can be replaced without significant architectural changes. Deliverables The completed project should include: Production-Ready Code • Python implementation • Modular architecture • Well documented • Unit tested where applicable Research Report Comprehensive report covering: • Methodology • Model comparison • Selection rationale • Validation process • Performance analysis • Parameter sensitivity • Limitations Backtesting Results Comparison between: • Static portfolio weighting • Dynamic regime-based weighting Including complete portfolio analytics and risk metrics. Documentation • Installation guide • Configuration guide • Model assumptions • Maintenance instructions • Extension guidelines Knowledge Transfer • Walkthrough session (or equivalent documentation) • Explanation of architecture • Future enhancement recommendations Ideal Candidate We are looking for a quantitative researcher/developer with experience in systematic trading and portfolio construction. Preferred qualifications include: • Strong background in quantitative finance • Experience developing systematic trading models • Practical experience with volatility forecasting • Experience with regime-switching models • Strong understanding of portfolio construction and risk management • Familiarity with multi-strategy trading systems • Experience working with futures and/or global financial markets Technical expertise should include: • Python • pandas • NumPy • statsmodels • arch • scikit-learn • hmmlearn (or equivalent) • Time-series modelling • Backtesting frameworks The ideal candidate should also be able to explain model choices, assumptions, and trade-offs clearly rather than delivering an opaque "black-box" solution. Project Outcome At the conclusion of the engagement, we expect a reusable portfolio intelligence framework that improves risk-adjusted returns by: • Dynamically forecasting market volatility for position sizing. • Identifying prevailing market regimes. • Adapting capital allocation between Trend Following and Mean Reversion sleeves. • Reducing concentration risk during periods when both sleeves become highly correlated. • Providing a modular foundation that can be extended with additional volatility models, regime classifiers, and portfolio optimization techniques as the systematic trading platform evolves.