Martin Fürholz — CTO

I built one of the first dedicated AI Red-Teaming operations at a major global testing company. My team conducted safety and security assessments on unreleased frontier models from leading AI labs. Before the public ever saw them. There was no playbook, so I wrote one. 60+ pages of methodology covering jailbreak taxonomies, LLM safety test cases, bias detection frameworks, and repeatable test cycle design. I trained testers and management alike, and served as lead operator and advisor on high-profile assessments for some of the biggest names in AI. THE TRACK RECORD This grew out of 7+ years as the sole technical anchor for the same company's Security Practice. Leading pentesting engagements across IoT, DRM, and critical infrastructure. I've personally triaged and verified tens of thousands of vulnerability reports and served as the final escalation point when standard teams hit their limits. 1,500+ verified vulnerabilities reported, including 28 critical findings in a single US banking subnet in a recent engagement. I also develop and operate custom agentic AI tools for autonomous reconnaissance. Currently delivering dark web intelligence and DRM piracy threat analysis to the C-level security division of a global Top-5 film studio. CURRENT FOCUS European digital identity infrastructure: I engineered the eID AI Security Scan, an agentic AI pentesting tool purpose-built for the EUDI Wallet reference architecture. It runs hybrid SAST/DAST mapped against eIDAS 2.0 ARF, BSI TR-03124, and OIDC4VP standards, with strict local data processing and zero cloud dependency. Agentic AI systems that replace entire multi-disciplinary team workflows, typically built and deployed in days, not months. I've done this repeatedly in production environments, not as prototypes. WHAT I OFFER AS AN INDEPENDENT CONSULTANT → AI Red-Team assessments: LLMs, agentic systems, RAG pipelines (engagements typically start at 5 days) → Security architecture reviews for AI-integrated products and government infrastructure → Agentic AI development: autonomous testing, reconnaissance, and workflow automation systems → Building your internal AI Red-Teaming capability from zero (strategy, methodology, team training) → eIDAS / EUDI-Wallet security assessments and compliance testing Full project portfolio and technical deep-dives: https://fuerholz.org If you're deploying AI or building critical infrastructure and nobody has tried to break it yet. That's not a question of if, but when. DM me or reach out at office@fuerholz.org.

Stackforce AI infers this person is a Cybersecurity expert specializing in AI Red-Teaming and security architecture.

Location: Rostock, Mecklenburg-West Pomerania, Germany

Experience: 0 mo