AI Engineering Portfolio
Building AI Systems for High-Stakes Environments
I’m a project and program management professional transitioning from 20+ years in emergency management and CBRN operations into AI systems architecture and cloud platform design.
This portfolio documents my journey building production-oriented AI platforms — systems designed not just to work in demos, but to operate reliably under real-world constraints.
What Makes This Portfolio Different
My approach to AI engineering is shaped by decades of leading complex, high-stakes programs where failure has consequences. I don’t just build models — I design the full platforms that support them:
- End-to-end system architecture — from data pipelines to service layers to deployment
- Operational readiness — reliability, monitoring, failure modes, and recovery
- Real-world constraints — security, scalability, maintainability, and cost tradeoffs
Technical Focus Areas
| AI & LLM Systems | Platform & Infrastructure |
|---|---|
| • LLM application architecture | • Backend service architecture |
| • RAG system design | • API design for AI workflows |
| • Multi-step orchestration | • Cloud-native AWS systems |
| • Model evaluation & monitoring | • Data platform design |
Core Technologies: Python • SQL • MongoDB • AWS • Git • CI/CD
Featured Projects
🚨 ChIRPS: Chemical Incident Response Platform Simulation
A web-based training platform for multi-agency CBRN emergency response exercises.
The Challenge: Traditional tabletop exercises lack auditability, time-ordering, and realistic multi-agency coordination dynamics.
The System: Full-stack platform enabling facilitators to run realistic incident simulations while capturing every decision, action, and communication in a queryable, time-ordered format.
Architecture Focus:
- Multi-user real-time coordination workflows
- Event-sourced data model for replay and analysis
- Role-based access and agency separation
- Automated After Action Report generation
📋 ERES: Emergency Response & Evaluation System
A local-first RAG pipeline that makes 500+ page Emergency Operations Plans instantly searchable during active incidents.
The Problem: Critical guidance is buried in massive PDFs, making it nearly impossible to find what you need when seconds count.
The Solution: High-performance retrieval system delivering grounded, source-cited answers to field responders.
Architecture Focus:
- Document chunking and embedding strategies
- Hybrid retrieval (vector + keyword)
- Citation and source attribution
- Local-first design for network-degraded environments
Portfolio Structure
This site is organized to support different levels of depth:
📍 Start Here (this page) → High-level overview
📂 Projects → Platform case studies with architectural deep-dives
🏗️ Architecture → System design patterns and tradeoff analyses
👤 About → Background, capabilities, and leadership philosophy
Why I’m Building in Public
I’m documenting this transition to:
- Demonstrate technical growth through hands-on system design
- Show architectural thinking — not just code, but design decisions and tradeoffs
- Build credibility in AI engineering through real, deployable systems
- Connect with others working at the intersection of AI and complex operational domains
Background Snapshot
From: Emergency management • CBRN operations • Federal program leadership
To: AI systems architecture • Cloud platform design • Production AI engineering
Focus: Systems that are reliable, maintainable, and suitable for high-stakes environments
Interested in AI System Architecture?