Things I've built, things I'm building, and things that are coming.
Problem: K–12 transportation programs lacked affordable, real-time visibility into fleet operations and student safety.
Solution: Built a multi-tenant SaaS platform with real-time GPS tracking, live map visualization, and white-label support for vendor isolation. Designed the architecture to be sector-agnostic for expansion beyond student transport.
Impact: Significantly reduced GPS tracking costs, enabled real-time fleet monitoring for school districts, and laid the foundation for a scalable multi-sector platform.
Problem: Requirements-to-code generation and engineering document creation are manual, error-prone, and lack traceability.
Solution: Built AI pipelines that automate code generation across multiple languages and engineering document creation with built-in compliance and auditability. Deployable across air-gapped, networked, and cloud environments.
Impact: End-to-end automation from requirements to tested, validated output — with full traceability and audit readiness.
Problem: SDLC tasks — requirements, architecture, code generation, testing, review, debugging — are disconnected and lack traceability across the pipeline.
Solution: Building an AI agent platform where specialized agents collaborate across the SDLC — with quality gates that reject bad work, audit trails for every decision, and human override when the AI can't figure it out. Configurable for different environments and industries. My own IP, built from scratch.
Impact: Configurable across aerospace, finance, healthcare, industrial, and consulting. Full traceability from requirement to deployed artifact. Name reveal coming soon.
Problem: Business rules are scattered across Jira, wikis, code, and people's heads. When sources disagree, nobody knows which one is right.
Solution: A knowledge platform that pulls from every source, surfaces conflicts when rules disagree, and lets humans establish the truth. Code is not the source of truth — it's evidence. Knowledge graph, conflict detection, business rules extraction.
Impact: The thing I wish I had on day one at every new company. Own IP, built from scratch. Name reveal coming soon.
Problem: Existing GCS software is vendor-locked, single-domain, or carries risky licenses for commercial or defense use.
Solution: Civilian, license-clean GCS for unmanned systems — air, ground, surface, subsurface. Hexagonal architecture, Apache-2.0, air-gapped capable. Domain-driven design with typed safety gates.
Impact: Multi-domain from day one. Dependency allowlist in CI. Own IP, built from scratch. Name reveal coming soon.
Problem: Drone platforms are tightly coupled to specific hardware. RF signal analysis tools are disconnected from system-level validation.
Solution: Device-agnostic framework for flight control, telemetry, and spectrum analysis. Single interface contract across devices and simulators.
Goal: AI-powered chatbot for this portfolio — answer questions about my experience, projects, and skills using RAG over my own content.