Research Vision

Agentic AI represents the next frontier in artificial intelligence—systems that don't just respond to queries but autonomously pursue goals, make complex decisions, and coordinate with other agents. My research explores the theoretical foundations, practical implementations, and societal implications of these transformative technologies.

Unlike traditional AI systems that operate reactively, agentic AI exhibits intentionality, planning capabilities, and the ability to navigate uncertain environments. This paradigm shift has profound implications for finance, healthcare, logistics, scientific discovery, and beyond.

Research Themes

01

Autonomous Agent Architecture

Designing cognitive architectures that enable agents to perceive, reason, plan, and act in complex environments.

  • Perception and world modeling
  • Goal-directed reasoning systems
  • Planning under uncertainty
  • Action selection and execution
  • Memory and knowledge representation
  • Meta-cognition and self-improvement
02

Multi-Agent Coordination

Understanding how multiple AI agents interact, compete, and cooperate to achieve individual and collective objectives.

  • Multi-agent reinforcement learning (MARL)
  • Cooperative and competitive dynamics
  • Communication protocols between agents
  • Emergent behavior in agent societies
  • Coalition formation and negotiation
  • Mechanism design for agent systems
03

LLM-Based Agents

Leveraging large language models as the foundation for sophisticated agentic systems with broad capabilities.

  • Tool use and API integration
  • Prompt engineering for agency
  • Chain-of-thought and tree-of-thought reasoning
  • Long-term memory and context management
  • Multi-step task decomposition
  • LLM agent evaluation frameworks
04

AI Agents in Finance

Deploying autonomous agents for trading, portfolio management, risk assessment, and financial decision-making.

  • Autonomous trading strategies
  • Agent-based market simulation
  • Strategic interaction modeling
  • Risk management agents
  • Sentiment analysis and information extraction
  • Regulatory compliance automation
05

AI Safety & Alignment

Ensuring agentic AI systems remain controllable, interpretable, and aligned with human values and intentions.

  • Value alignment and reward modeling
  • Corrigibility and shutdown protocols
  • Interpretability and explainability
  • Robustness to adversarial inputs
  • Ethical constraints and guardrails
  • Human oversight mechanisms
06

Human-AI Collaboration

Designing systems where AI agents augment human capabilities while respecting human autonomy and judgment.

  • Collaborative decision-making frameworks
  • Delegation and task allocation
  • Explainable recommendations
  • Trust calibration and transparency
  • Interface design for agent systems
  • Skill complementarity analysis

Key Publications in Agentic AI

2025

Autonomous Agents in Decentralized Finance: Coordination and Emergent Behavior

Journal of Financial Economics (forthcoming)

Examines strategic interaction between AI trading agents in DeFi protocols, documenting emergent market dynamics and coordination failures.

2024

Agentic AI: From Theory to Practice

MIT Press

Comprehensive 600-page textbook covering foundations, architectures, learning algorithms, applications, and safety considerations for autonomous agents.

2024

Large Language Models as Market Participants

Review of Financial Studies

First empirical study of LLM-based trading agents, analyzing their strategies, coordination patterns, and market impact across asset classes.

2024

Multi-Agent Reinforcement Learning: Theory and Applications

Artificial Intelligence Journal

Theoretical analysis of convergence properties in MARL systems with applications to cooperative and competitive environments.

2023

Explainable AI Agents: Transparency in Autonomous Decision-Making

Proceedings of NeurIPS

Novel framework for generating human-interpretable explanations of agent behavior in complex sequential decision-making tasks.

2023

Value Alignment in Autonomous Trading Systems

Journal of AI Research

Addresses AI safety challenges in financial applications, proposing methods to align agent objectives with firm goals and regulatory requirements.

2023

Emergent Communication in Multi-Agent Environments

International Conference on Machine Learning (ICML)

Studies how agents develop communication protocols to coordinate behavior, with implications for AI interpretability and control.

2022

Tool-Using Language Model Agents: Capabilities and Limitations

Proceedings of AAAI

Systematic evaluation of LLMs' ability to use external tools, identifying failure modes and proposing architectural improvements.

Real-World Impact

My agentic AI research has direct applications across multiple industries:

Financial Services

Autonomous trading systems managing portfolios, executing complex strategies, and providing real-time risk assessment for major hedge funds and proprietary trading firms.

Healthcare

AI agents for treatment planning, resource allocation, and clinical decision support, collaborating with physicians to improve patient outcomes.

Supply Chain

Multi-agent systems coordinating logistics, inventory management, and procurement across complex global supply networks.

Scientific Discovery

Autonomous research agents designing experiments, analyzing results, and generating hypotheses in drug discovery and materials science.

Graduate Courses

Agentic AI Systems (Ph.D. Seminar)

Advanced seminar covering theoretical foundations and cutting-edge research in autonomous agents. Topics include cognitive architectures, MARL, LLM agents, safety, and alignment. Students complete original research projects.

AI for Business Applications (MBA Core)

Introduction to AI agent systems for business students. Covers practical applications in strategy, operations, and decision-making with hands-on projects building AI-powered business solutions.

Multi-Agent Systems (MS/Ph.D. Elective)

Technical course on multi-agent coordination, game theory, mechanism design, and emergent behavior. Includes implementation projects and agent-based simulations.

Collaborate on Agentic AI Research

Seeking research partnerships, Ph.D. student collaborations, and consulting opportunities in autonomous agent systems

Get in Touch