Steve Becht-Buss AI Cred Profile

30+ years enterprise data architecture. Now building production GenAI-RAG systems on relational databases — no separate vector database needed. 5NF design means the database enforces referential integrity so LLM code-generation tools produce dramatically better code. Works on any relational database.

Built a multi-tenant RDBMS-hosted RAG platform — 120,000 lines of code in four months, 100% AI-generated. SQL Server, Claude Code. Adoption across business units has begun.

Also: AI governance frameworks (information classification + Microsoft Purview enforcement), legacy app refactoring with Claude Code, SDLC document generation, and open-source Mermaid Diagram Guidelines (MDG) for structured thinking with LLMs.

AI CRED Score: 8.7 "High Velocity" | 9.5/10 Practical Application

Steve Becht-Buss is an Enterprise AI/Data Architect with 30+ years of experience designing production-grade systems at the intersection of relational database architecture and generative AI. Operating at expert-level AI fluency, Steve builds what few others attempt: production RAG systems hosted entirely within relational databases, eliminating vector database dependencies through 5NF design principles that dramatically improve LLM code-generation quality. He delivered a 120,000-line, AI-generated multi-tenant RAG platform in four months — now expanding across business units. His workflow architecture is distinctive: four-file context engineering, versioned skill files, commit-hook-enforced verification, and multi-model review pipelines that treat convergent criticism as signal. Steve applies Deutschian fallibilism as a design principle, building self-correction mechanisms into every system rather than trusting outputs. He brings organizations a rare combination — deep architectural judgment, transferable methodology through his GenAI Empowerment Program, and a practitioner actively working at the frontier of human-AI collaboration.

Steve Becht-Buss

About

AI Wingman Summary

Links

Get Your AI Fluency Score