Agentic systems, distributed platforms and secure backends — from architecture to operations. Evidence, evaluation and observability over demos.
Three proof areas up front: AI / Agent Systems, Product / Full-Stack, and Backend / Security & Distributed Systems — each with ownership, decisions, trade-offs and evidence.
A full-stack restaurant ordering and operations platform: Next.js frontend, ASP.NET Core API, PostgreSQL, SignalR live kitchen updates, role-based dashboards, Docker/DevContainer setup, CI and documented architecture.
ARCHITECTURE
A reproducible AI-engineering reference system built across eight maturity stages: risk/product contracts, retrieval, evaluation, provenance, approvals, observability, security and progressive delivery without provider lock-in.
ARCHITECTURE
A production-grade, modular auth backend: OTP two-factor login, granular RBAC, token rotation and waitlist management — built as a self-hostable foundation for SaaS platforms instead of another vendor-locked BaaS dependency.
ARCHITECTURE
Production-ready applications from architecture through deployment. Typed frontends, real state management, and interfaces that stay maintainable past the first release.
Production AI systems beyond prompt wrappers: retrieval pipelines, agent orchestration, evaluation harnesses and observability for systems that must behave predictably.
APIs, services and data layers that scale without becoming unmaintainable. Security-first defaults, explicit trade-offs, infrastructure you can actually operate.
Constraints before code. Domain, existing systems, failure modes and real requirements — understood on paper first.
Architecture before implementation. Data model, interfaces and trade-offs made explicit and documented.
Incremental vertical slices. Working software early, tested as it grows — not a big reveal at the end.
Automated pipelines, observability and rollback paths from day one. Deployment is part of the design.
Production ownership. Monitor, measure, maintain — the system has to keep working after launch.
The same discipline whether the engagement is a four-week build or a year-long system: understand constraints first, design before implementing, ship incrementally, and keep owning what runs in production.
Grade 99/100 — full-stack product work with architecture, implementation, testing and project documentation under a fixed assessment scope.
View project ↗Research papers, case studies, API notes, architecture documentation and operating notes kept alongside the systems they describe.
Browse writing ↗Restaurant Ordering System, UserMatrix, Orion, HomeLab and SSH-Hack are public, inspectable codebases with architecture or operational documentation.
View on GitHub ↗Book and research track around SupportPilot, Graph Engineering and controlled agent systems: contracts, evidence, evaluation, approvals and release gates.
Read research ↗Released the practical AI-engineering book and companion system track: risk contracts, retrieval, evaluation gates, provenance, approvals, observability and progressive delivery.
Read dossier ↗Public Next.js + ASP.NET Core platform with PostgreSQL, SignalR, role dashboards, Docker/DevContainer setup, GitHub Actions CI and documented architecture.
View repo ↗35-section position paper on building production-grade agent systems: graph topology, contracts, verification, observability, and cost models.
Read paper ↗Zero-trust coordination layer for multi-agent systems with selective transport gates, polarization state channels, and eBPF/TrustZone hardware acceleration. Go + Rust reference implementations with benchmarks.
Read paper ↗Production-ready auth foundation: OTP 2FA, RBAC matrix, JWT rotation, waitlist management. TypeScript/Node, PostgreSQL, Redis. Self-hostable, containerized.
View on GitHub ↗AI Systems Engineer based in Germany — agentic systems, distributed platforms and secure production engineering. Full-stack and backend depth as foundation, not headline.
I take ownership end to end: understanding the domain, designing the architecture, shipping incrementally and operating what ships. The rest is in the dossier.
Tell me what you're building, what's blocking you, and where you need engineering help.
