Anjan Roy

Agentic AI &

Portfolio Decision Systems

Portfolio Management as an Evolving Decision System

I apply Agentic AI and rule-based architectures to financial decision environments. My work focuses on the structural reality of markets: non-stationarity, assumption decay, and the necessity of rigorous feedback loops.

The lens is precise: portfolios, derivatives, and systemic risk. Unlike prediction-focused models that attempt to forecast price, decision systems focus on response, control, and adaptation across shifting regimes.

Scope of Practice

  • Decision Systems Architecture
  • Derivative Risk Control
  • Regime-Aware Portfolio Logic
  • Human-in-the-Loop Governance

* Note: This work is distinct from signal generation, high-frequency trading, or consumer products.

Areas of Research & Practice

Agentic AI & Decision Systems

  • Agentic and rule-based system design
  • Monitoring and feedback architecture
  • Constraints and effective guardrails
  • Observability and decision diagnostics
  • Human-in-the-loop governance models

Portfolios, Derivatives & Risk

  • Portfolios as dynamic decision systems
  • Derivatives as control instruments
  • Volatility, convexity, and payoffs
  • Risk as system instability, not point estimates
  • Structuring for survival and compounding

Market Reality & Regimes

  • Managing assumption decay
  • Detecting and adapting to regime shifts
  • Non-stationarity in financial data
  • Divergence of theory from practice
  • Fragility of over-optimized models

Program Formats

Executive / Faculty Capsule

Duration: 2–3 Hours

A strategic session designed for faculty alignment and program anchoring. Focuses on why classical assumptions break down in modern markets and how agentic frameworks provide decision discipline.

Key Themes

  • • Breakdown of classical assumptions
  • • Agentic AI for decision discipline
  • • Realistic derivative usage
  • • AI value attribution & limits

Agentic AI for Portfolio & Derivatives Decision Systems

Core Immersion • 3 Days (Practitioner-Led)

A deep-dive immersion designed to complement existing finance, quant, or systems curricula. Offline-first, rigorous, and pilot-friendly.

Day 1: Decision Systems

Portfolios as systems, assumption decay, monitoring vs. prediction. Introduction to agents, feedback loops, and human-in-the-loop governance.

Day 2: Risk Limits

Volatility and convexity. Using derivatives to shape risk and stabilize portfolios across regimes. Practical payoff walkthroughs (conceptual).

Day 3: Agentic Integration

Mapping decision rules, monitoring across regimes, risk as system behavior. Failure modes, diagnostics, and responsible deployment.

+91 999 062 0500(Mobile & WhatsApp)
© 2026 Anjan Roy