Skip to content

AI consulting with technical substance

I help companies identify useful AI use cases, assess technical feasibility realistically, and turn that into clear roadmaps, architecture decisions, or pilot plans.

How I can help

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.

02

AI Architecture Review

I review AI ideas, prototypes, or planned integrations from an architecture perspective: data flows, RAG, agents, APIs, fallbacks, security, and maintainability.

03

AI roadmap and pilot planning

I translate an AI idea into a realistic pilot plan with requirements, data needs, risks, milestones, and success criteria.

Fit

For technical teams and product owners

When AI should be integrated into processes, products, or internal tools and architecture, data reality, and operations need realistic assessment.

Boundary

Not every task needs AI

Some problems first need better processes, better data, or classical automation. Good AI consulting also identifies when AI is not the right first step.

How I work

  1. understand context, goal, and users
  2. review use cases, data situation, and existing systems
  3. assess technical options and boundaries
  4. make risks, stop criteria, and open decisions visible
  5. document next steps as a roadmap, architecture proposal, or pilot plan

Project evidence

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

BetriebsGehirn

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

Speech Therapy Report Agent

AI-assisted tool for structured speech therapy reports

Transcription, structured extraction, and validated draft reports with FastAPI/Pydantic.

Live system

This Portfolio

Live proof of AI/LLM integration

Prototype

CoWork-OS

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.

View project evidence →

AI & Privacy: AI features may send inputs to Google Gemini, OpenAI, or Groq. Do not enter sensitive data.

Details