I design serverless-first cloud systems for startups and scale-ups. My own production platform runs on $8.70/month. Every recommendation I make, I've already deployed and watched under load.
Built to run. Not to convince.
Production-grade architecture, not slide decks. Every engagement ends with working infrastructure or a migration-ready blueprint.
Workers, Pages, Durable Objects, R2, WAF. Zero-origin security patterns. Edge-native design that eliminates your origin server.
Fargate Spot, Lambda, EventBridge, cost optimization. Eliminate idle compute. ARM64 migration for 40% cost reduction.
Cross-provider audit, vendor lock-in analysis, cost arbitrage. Identify which workloads belong where and why.
CI/CD pipeline security, OWASP compliance, secret management. Full supply chain review from commit to production.
Cluster design, Helm chart strategy, horizontal scaling, cost optimization. Right-sized infrastructure for your actual workload.
Three phases. No ambiguity about what you get or when.
60-minute call to map your current architecture, constraints, and goals. Fixed-price quote within 24 hours. No surprises.
Deep-dive into your infrastructure, codebase, and cost structure. AI-enhanced pattern recognition accelerates the audit. Every finding includes evidence.
Architecture document, migration roadmap, and implementation guide. Optional hands-on deployment. You own everything.
PairSignal is a cointegration-based statistical arbitrage platform I built and operate. It runs 24/7 with 8 defense layers, 1,256 automated tests, and a total infrastructure cost under $9/month. This is not a demo. It trades real capital.
I run an AI-native architecture advisory practice based in Germany. I design serverless-first cloud systems, conduct DevSecOps audits, and build multi-cloud strategies for startups and scale-ups.
Languages: English (fluent), Spanish (native).
I use AI-enhanced methodology (Claude Opus 4.6) to accelerate pattern recognition and ensure thoroughness across large codebases and infrastructure configurations. The judgment and recommendations are mine; the AI handles research velocity.
This advisory practice uses AI tools (Anthropic Claude) as a research and analysis accelerator. AI assists with pattern recognition, documentation review, and code analysis. All architectural decisions, recommendations, and deliverables are made and verified by a human practitioner. AI-generated content in deliverables is disclosed. No autonomous decision-making is delegated to AI systems.
Describe your architecture challenge. I respond within 24 hours with an honest assessment of whether I can help.
Your documents and infrastructure details are processed confidentially and never stored beyond the engagement. We retain nothing. All work is performed under NDA upon request.