The Prompt Decalogue
Universal rules for talking to LLMs that strip out ambiguity at the input stage.
I advise on, design, and oversee AI deployments for companies and local governments — with a hard focus on quality, safety, and eliminating hallucinations. Decisions, architecture, communication — and accountability for the result.
Deploying AI isn't prompt engineering — it's architecture that prevents errors. Most models work fine in a proof of concept, then fall apart in production. My kernel changes that. Full demonstration of the method: the experiment "One Question, Split Across 91 Minds" (in Polish).
Universal rules for talking to LLMs that strip out ambiguity at the input stage.
An architecture in which independent models — Gemini, GPT, Claude — cross-check each other's output.
Version-controlled AI context that enables a precise rollback to a stable state.
Feedback loops that automatically improve answer quality based on the data.

A 3-in-1 device designed by AI. Modular ("Lego-style") measurement units, enclosure, and architecture, generated iteratively by LLMs. Hardware cost cut from €10,000 to €20.

A dedicated B2B service portal for electronics manufacturers. RMA automation, duplicate detection (fraud prevention), and real-time SLA tracking. Zero manual paperwork.

An enterprise-grade system for Polish municipalities. 100% offline PWA, full WCAG 2.1 AA compliance, 4 languages (PL/EN/DE/UA). Deliberately framework-free — runs on aging municipal servers, easy to audit.

Monitoring of Schumann resonance, solar activity, and the geomagnetic field from 6 stations worldwide — AI translates raw data into plain language. No accounts, privacy-first, an open API built for AI agents (llms.txt, RFC 9727).
A free biorhythm app based on Dr. Sikora's method — 8 languages, 288 articles, zero accounts. I'm running an open study of the method: a blinded N-of-1 journal (Pearson correlations, effect size) plus an anonymous opt-in aggregate. I'm checking honestly whether the method works — instead of just selling it.
The biorhythm engine as a public npm package, v2.0 (MIT): 952 golden tests, validated digit by digit against the original 1983 method. Code anyone can check — that's what "AI that doesn't lie" looks like in practice.
Self-hosted, cookie-free analytics (Plausible CE) for ~20 domains. Visitor privacy plus full data ownership.
Engagement models, pricing, and retainer (fractional CTO).
View →One of the first cases in Poland of AI acting as a campaign's lead strategist. Enterprise-level capability on a local budget: communication strategy, content generation, live sentiment analysis.
Designing and deploying enterprise-grade AI systems. Focus: multi-model orchestration, hallucination elimination (AI Quality Loop), token cost optimization.
Full digital transformation of a manufacturing company. 200% growth in online reach. Rolled out the M2 Service Portal (90% cut in support costs).
Co-founded and scaled one of Poland's first cloud hosting providers (top 5 nationally). Grew the customer base from zero to 15,000+ B2B companies.
"Technology without communication is a cost. Communication without technology is a promise."
2003–2008 · Thesis: "Growth Strategies for Small and Medium Enterprises"
Strategy, audit, or implementation — reach out through the form and I'll get back to you with specifics.
Contact form