01
Identifying useful AI opportunities
I assess processes, knowledge flows, and use cases to separate meaningful AI opportunities from classical automation or data-quality work.
I help companies identify useful AI use cases, assess technical feasibility realistically, and turn that into clear roadmaps, architecture decisions, or pilot plans.
01
I assess processes, knowledge flows, and use cases to separate meaningful AI opportunities from classical automation or data-quality work.
02
I review AI ideas, prototypes, or planned integrations from an architecture perspective: data flows, RAG, agents, APIs, fallbacks, security, and maintainability.
03
I translate an AI idea into a realistic pilot plan with requirements, data needs, risks, milestones, and success criteria.
Fit
When AI should be integrated into processes, products, or internal tools and architecture, data reality, and operations need realistic assessment.
Boundary
Some problems first need better processes, better data, or classical automation. Good AI consulting also identifies when AI is not the right first step.
Selected own projects show how I assess AI use cases technically, embed them into workflows, and connect them with architecture, security, and quality boundaries.
In development
Local knowledge infrastructure with RAG, source status, and answer confidence
RAG is more than vector search here: search scope, source grounding, evidence status, structured output logic, and honest non-answers. Local architecture prototype, not a finished product.
Read the case note →Prototype
AI-assisted tool for structured speech therapy reports
Transcription, structured extraction, and validated draft reports with FastAPI/Pydantic.
Live system
Live proof of AI/LLM integration
Prototype
Markdown control plane for agentic software development
Tool-agnostic workflows, project truth, handoffs, and review gates for Codex, Claude Code, Gemini, and other AI tools.