I am Wojciech Wiesner — former Founder & CEO of Xpress Delivery. I design and operate AI products where deterministic workflow logic, validation and human accountability define the boundary; models handle bounded interpretation, extraction and language.
Across these named systems, deterministic workflow logic governs business decisions; models are used for bounded interpretation, extraction and phrasing.
Borg = many collapsed into one. The Ones inverts it: one person, amplified into many — with the human owning every decision and every quality gate.
Human intent
You set the outcome. The operator stays sovereign — never assimilated.
Orchestrator
Intent is decomposed into scoped work a specialized team can execute.
Specialized agents
A voluntary collective of AI agents does the parallel heavy lifting.
Quality gates
Tests, reviews and scans provide evidence before release.
Verified result
Output with evidence attached — not a promise, a track record.
Your CV says what you claim. Your code contains evidence. Developer Passport connects the two — evidence-first, with confidence levels and an explicit account of what it cannot infer. No automated verdict on whether you are a “good” developer.
Founding cohort is a consent-first waitlist. No GitHub token or private-repo access in this phase.
A personal toolkit and lab, not one finished SaaS: a reproducible portfolio audit, self-hosted model routing, coding agents andBorg Cockpit(live). Many agents and tools, coordinated as one system.
Before building AI products, I founded and led Xpress Delivery, a same-day courier business spanning product, engineering and enterprise B2B operations. In 2022, Brief included me in “50 Most Creative in Business”. That operating background is why I start with the business bottleneck, the decision boundary and the outcome to measure — not the model.
Brief: “50 Most Creative in Business” (2022)A/B benchmark na 5 modelach (w tym Llama 3.1 8B lokalnie na Mac mini). Prompt tokens spadły o 59-88%, latencja do 58%. Zero ściemy: jakość była identyczna.
Read article →GAIA L1, 53 tasks: harness fixes took us from 34 to 45 correct. A conscience gate and a judge added two more, at 1.7x the time.
Read article →No JIT, JIT without JEV, and JIT with a live JEV call on the same 10 fixtures. Wall clock fell to 1.57x. Tokens did not.
Read article →I work from operational problem to governed delivery: clarify the decision, design the workflow boundary, lead the build, and verify the result in the real system.