Navigating the Balance Between Model Complexity & Interpretability with BNY Mellon
Spotlight Interview: Rohit Sharma, Vice President, Market Surveillance at BNY Mellon Investment Management
What’s the most exciting opportunity you see right now in quant strategy or data-driven investing?
The most exciting opportunity today lies in the convergence of AI/ML with traditional quantitative strategies, especially in the realm of real-time trade surveillance and adaptive execution models. The ability to model market microstructure with deep neural networks and enhance trading decisions through dynamic reinforcement learning is fundamentally reshaping how quants interact with markets. Tools like SHAP and LLM-based explainability are also enabling greater trust in models — a game changer for regulatory and compliance-driven environments.
What’s one major challenge or disruption quants need to solve in the next 12–24 months?
One of the biggest challenges is navigating the balance between model complexity and interpretability, particularly with the increasing use of deep learning in regulated financial environments. As models become more opaque, compliance expectations around transparency and auditability are rising. Bridging this gap without compromising performance will be key. Additionally, data quality, latency, and security remain critical bottlenecks, especially as firms scale cross-asset, cross-venue platforms.
What are you most looking forward to discussing or hearing about at Quant Strats EU EXPO 2025?
I’m most looking forward to discussing the application of deep learning in algorithmic execution and the evolving architecture of real-time trading and surveillance systems. I’m also excited to hear diverse viewpoints from peers tackling AI governance, regulatory technology, and efficient infra for low-latency analytics — themes that are shaping the next wave of quant innovation.
To connect with Quant leaders like Rohit Sharma this October book your pass today.
