With software, AI, automation or infrastructure. Without more complexity than necessary.
I work directly with small, decisive companies. Ideally with people who explore AI themselves, can make decisions quickly and want someone who understands the whole problem, builds the solution and stays around when it runs in production.
Generalist in capability. Opinionated in method.
I do not fit neatly into developer, AI consultant or DevOps. Good. Business problems do not respect departments either. I can understand the workflow, build the software, use AI where it helps and take care of operations.
Complexity is a cost, not proof of competence.
I build the smallest system that reliably does real work. AI where language, context or uncertainty belong. Ordinary software where explicit rules are better. Solid infrastructure underneath so the demo still works on Monday morning.
I do not want ticket number 847. I speak directly with the person who owns the outcome, understand the workflow and solve the bottleneck end to end.
Fewer tools, handoffs and architecture theater. As simple as possible, as robust as necessary and understandable to the people using it.
I can advise, code, automate and own infrastructure. No agency telephone game and no disappearing after the presentation.
Real workflows instead of AI theater.
Different industries, one pattern: understand a real problem, remove unnecessary complexity and build a system that holds up in daily work. The range is intentional.
I help move AI voice agents out of the demo corner and into real hotel processes: understanding requests, checking availability, respecting location-specific rules and handing off cleanly to existing systems.
I work on the next evolution of the platform: from individual digital processes toward agent-first workflows, better handoffs between portal, admin and ERP, and groundwork for partially autonomous operations.
I support the move from a grown content and membership setup to a dedicated platform: less tool chaos, better member processes, more self-service and room for new digital products.
I help automate operations across customer service, accounting and marketing. Not “one bot here, one Zap there,” but rebuilding processes so recurring work runs cleanly through systems.
I showed how AI can be used practically for construction documentation and tender work: reading plans faster, preparing bills of quantities and structuring site reports from voice.
What started as process automation became a whole product line. I helped shape how recurring evaluations and manual workflows turn into an engine that drives the startup’s product.
AI becomes useful when it is attached to genuinely annoying workflows.
Most companies do not need a giant AI master plan. They need one first workflow that visibly removes work without pretending everything should run autonomously.
Today: Humans read every request, search context and write similar replies.
First version: AI classifies, prioritizes and drafts responses.
Human stays in: difficult cases, goodwill, legal statements.
Today: Data is collected, copied, explained and reformatted every week.
First version: Collect, structure, comment and output it as a repeatable report.
Human stays in: judgement, decision, final approval.
Today: Calls interrupt the team and important information lands somewhere unstructured.
First version: A voice agent captures data, checks simple rules and hands off cleanly.
Human stays in: edge cases, sales, sensitive conversations.
Today: Plans, invoices, contracts or tenders are read and transferred manually.
First version: AI extracts structure, marks uncertainty and prepares downstream processing.
Human stays in: expert review and liability decisions.
Today: Knowledge is scattered across heads, emails, docs, tickets and old projects.
First version: A searchable knowledge system with sources, examples and clear boundaries.
Human stays in: maintenance, prioritization and new decisions.
Today: Inbox, CRM, accounting, project management and customer data barely talk to each other.
First version: A workflow that passes data cleanly and only pulls humans in for exceptions.
Human stays in: control, edge cases and process improvement.
The first step can be one workflow. The relationship can grow beyond it.
The workshop is a concrete, low-risk door, not my whole identity. If we work well together, I can stay as a technical sparring partner, builder or operator.
30 minutes. I look at your workflow and tell you honestly where AI makes sense and where it does not.
Half-day 1:1 session. We build the first working workflow live.
For operators who regularly need to untangle problems, make decisions and keep building with AI or software.
More implementation and ownership. I build and improve software, automations, AI systems and the infrastructure underneath.
Building together, not for you.
We analyze your workflow end to end. Where does time disappear? What repeats? Where is the real leverage?
We build the solution live or in short iterations. Not abstract consulting, but a working system.
You leave with tools, understanding and next steps. The goal is capability, not dependency.
Do not claim what I can do. Show what I built and learned.
I write about active projects, decisions, failed attempts and the boundary between AI, ordinary software and human work. Some posts will be interactive: clickable workflows, real execution traces, small calculators and visual explanations.
Show me. I will tell you honestly whether I can help, whether AI belongs in the solution and what the simplest sensible next step is.