Portfolio Construction
& Optimization
Our convex optimization pipelines dynamically translate alpha scores into target positions while rigorously penalizing transaction costs and unwanted factor exposures. Rather than relying on static Markowitz frameworks, we utilize Hierarchical Risk Parity (HRP) and deep reinforcement learning (DRL) agents to solve high-dimensional, multi-period allocation problems. This ensures maximum capital efficiency and turnover minimization across thousands of traded equities daily.
Key Competencies:
- Deep Reinforcement Learning (DRL) for dynamic, multi-period portfolio trajectory optimization.
- Hierarchical Risk Parity (HRP) and graph-theory-based clustering for true diversification.
- Non-linear transaction cost modeling taking into account market impact and order book liquidity.