David “Dave” R. Rintoul, CISSP
AI Systems & Agentic Architectures
MCP & LangGraph Workflows: Agentic pipelines that pull live data so LLM answers stay grounded in real information, countering the three weaknesses of consumer AI: missing context, model hallucination, and stale training data.
A live example is Voyages By Dave, which combines weather, travel advisories, events, and vendor knowledge into one travel research query.
RAG Semantic Search
Turn scattered documents into attributable answers. This self-hosted RAG stack scrapes, chunks, embeds, and reranks content so you can query a private knowledge base with verified sources. The result: every answer cites the source document it came from, so teams can ground decisions in real evidence instead of repeated document hunts.
High-Stakes Governance & Safety
Operational Leadership: Directed compliance and risk management in healthcare and commercial diving, where a missed check has immediate consequences. That discipline informs AI governance under pressure, preventing avoidable mistakes in high-stakes systems.
AI Security Translation: Turn operational discipline into guardrails that keep systems secure and useful. tableau-xml-parse: a Python audit tool that traces Tableau calculated fields back to database columns, so governance can verify schema changes before they break reports.
Global Alliances & Revenue Scaling
Microsoft: Restructured the 200+ partner program for a customer care product, focusing engineering and field investment on strategic partners while maintaining broad ecosystem coverage.
Siemens: Equipped global sales teams and consolidated 3 telephony platforms into 1 while preserving customer functionality and reducing R&D costs. The result: scaled high-value alliances and reduced R&D cost by consolidating platforms, without losing customer functionality or ecosystem coverage.
Is It Canadian? Demo
An evidence-driven checker that vets whether a company is Canadian. An agentic workflow finds the official site, scrapes corporate pages, and a local model classifies from collected evidence. The result: compliance and procurement checks that used to take hours now finish in minutes, with a citation trail for every verdict.
Firecrawl Extended
A self-hosted Firecrawl deployment that fills the gaps. A single REST gateway fronts every upstream endpoint — scrape, crawl, map, search — plus a built-in web console, while a custom Playwright service adds browser sessions (clicking, form-filling, screenshots) behind /v1/interact, and /v1/extract runs on a local Ollama model — no external LLM calls. Agents reach the same surface via an MCP interface. The result: full-featured scraping and browser automation on your own infrastructure, not in third-party logs.