SOFTWARE ENGINEER · ENTERPRISE AI · BACKEND SYSTEMS

I build AI systems that hold up in the real world.

Computer scientist by training, software engineer by practice. I build reliable, provider-agnostic AI and backend systems for real enterprise workflows.

Educated in Germany and shaped by enterprise software, security, insurance, and quantum computing, I evaluate models by quality, latency, cost, privacy, and deployment constraints — not provider popularity.

View selected work→Let’s talk→

Stay steadfast. Do good. Pursue excellence.

“Build with real systems. Reality reveals what theory leaves out.”

4+ years engineering5,000+ monthly users35% lower latency40% higher throughputM.Sc. Computer Science · LMU

Reliable AI Systems

Evidence-bound RAG, agentic workflows, evaluation, observability, and explicit human-control boundaries.

⌘

Enterprise Engineering

Maintainable backend and event-driven systems with Java, Spring Boot, Python, FastAPI, Kafka, and PostgreSQL.

⇄

Model Optionality

Provider-agnostic architectures and task-specific model selection across quality, cost, privacy, and deployment constraints.

EXPERIENCE

Experience behind the systems.

The engineering contexts that shaped how I design and deliver real systems.

ENTERPRISE AI

Bringing AI into real workflows

Internal knowledge systems, evidence-bound RAG, agentic workflows, streaming, evaluation, and observability — built within real data, access, and operational constraints.

BACKEND & SCALE

Keeping systems stable under load and change

Event-driven services, APIs, databases, and distributed components with measurable improvements in latency, throughput, and maintainability.

SECURITY

Treating security as a system boundary

Authentication, RBAC, OAuth2, auditability, and OWASP-aligned engineering for systems with sensitive data and explicit accountability.

REGULATED DOMAINS

Engineering where failure has consequences

Experience across insurance, enterprise, and research environments informs my focus on traceability, human oversight, and controlled system behavior.

Selected Work

AI systems and software foundations designed around real workflows, explicit boundaries, and inspectable engineering decisions.

GitHub profile→

S1Flagship · In active development

SupportOne

Evidence-Bound Insurance Operations

An insurance agent that cites its evidence, refuses unsupported answers, and gates consequential actions behind application policy and human approval.

LangGraphFastAPIpgvectorNext.js