Agentic AI • Technical Program Leadership • Applied AI

Chantell Harris-Headley
Agentic AI Portfolio

A portfolio of hands-on AI systems demonstrating autonomous agents, multi-agent orchestration, Retrieval-Augmented Generation (RAG), tool use, workflow state management, evaluation concepts, privacy-oriented controls, and business-focused AI delivery.

Johns Hopkins University — Certificate Program in Agentic AI PythonLangGraphLangChainRAGChromaDBOpenAIAI Workflow Design

Featured Projects

These projects were developed to demonstrate how agentic AI can be applied to business research, decision support, and coordinated multi-agent workflows. Public demos use sanitized or simulated data where appropriate.

Autonomous Agent

Autonomous Financial Research Agent

A goal-oriented research agent that coordinates multiple tools to assemble evidence, analyze sentiment, retrieve relevant research, and synthesize a structured financial briefing.

  • Autonomous tool selection and orchestration
  • Market data, news, sentiment, and historical analysis
  • RAG and vector retrieval concepts
  • LangGraph state management and routing
  • Error resilience, confidence, and source transparency
Multi-Agent System

Multi-Agent Mortgage Underwriting

A coordinated underwriting workflow using a Supervisor and specialized agents for credit, income, assets, collateral, review, and final decision support.

  • Supervisor-based agent orchestration
  • Credit, Income, Asset, Collateral, Critic & Decision agents
  • RAG-based policy retrieval
  • Deterministic financial calculations
  • PII masking, risk scoring, and human-review escalation
RAG + Analytics

DualLens Analytics

A decision-support concept that combines quantitative financial performance with qualitative company AI-strategy research to evaluate both business strength and innovation potential.

  • Financial metrics and comparative analysis
  • RAG over company AI-strategy documents
  • Embeddings and semantic retrieval
  • Grounded LLM synthesis
  • Ranking and visualization of investment-oriented signals
Case Study

What These Projects Demonstrate

My focus is not only on making an agent run. I approach AI systems as programs that require clear business objectives, defined workflows, state and dependency management, measurable acceptance criteria, risk controls, testing, documentation, and stakeholder-ready communication.

Agent ArchitectureTool CallingMulti-Agent OrchestrationRAGPrompt DesignState ManagementError HandlingResponsible AIHuman-in-the-LoopEvaluation ConceptsTechnical DocumentationProgram Delivery

Portfolio Positioning

Target roles:
AI Program Manager
Technical Program Manager — AI
Agentic AI Program Lead
AI Transformation / Enablement Lead

Focus:
Bridging business objectives, technical teams, AI system design, governance, and delivery.