AI/ML Engineer for Autonomous Pentesting

Company DescriptionSQUR is building Europe’s leading fully autonomous pentesting platform, redefining how cybersecurity is delivered.Our mission is to help organizations — especially those with limited security resources — achieve continuous protection without waiting weeks for human testers.By combining AI agents, security expertise, and deterministic exploit automation, SQUR delivers faster, more consistent, and more transparent results than traditional approaches.We’re a lean, fast-moving team in Karlsruhe, Germany, backed by strong partners and focused on making autonomous security a European strength.Role DescriptionWe’re looking for a Founding AI/ML Engineer to design and build the intelligence behind SQUR’s autonomous pentesting agents. This is a full-time role based in Karlsruhe (preferred) or remote within Germany.ScopeOwn the agent system end-to-end: planning, tool arbitration, memory, recovery, and long-running job control.Build and operate ML pipelines for data curation, labeling, evaluation, training, and deployment.Fine-tune and integrate LLMs for reliable tool use, reasoning, and remediation guidance.Drive reliability at scale: retries, fallbacks, circuit breakers, stateful progress, and cost/latency optimization.Define verification signals (model confidence, evidence quality, chain validity) and feed them back into planning and action.Harden safety and isolation: sandboxing, tool isolation, and orchestration designed for real targets.Set engineering standards, telemetry, and evals; mentor the team toward measurable end-to-end outcomes.CollaborationPartner with offensive security researchers to translate exploit workflows into safe, deterministic agent behaviors, and with product to deliver compliance-ready outputs.Qualifications5+ years building production ML/agent/back-end systems with verifiable impact (open source, papers, talks, or shipped products). Familiarity with agent frameworks (e.g., LangGraph/LangChain) and structured evals for agents.Expert Python for production systems: orchestration, async/state machines, containers, packaging, and testing.LLM & agent expertise: tool-use/function-calling, planning, retrieval, and fine-tuning in production.Reliability & scale: resilient execution (retries/fallbacks), cost/latency optimization, observability, and SLIs/SLOs.Eligible to work in Germany (Karlsruhe preferred; remote in DE possible).Nice to have: AI-native mindset; experience with offensive security workflows or scanners (e.g., ZAP/Burp/Nuclei) is a plus.Experience with GCP/Kubernetes/Cloud Run, containers, queues, and distributed execution.What you getFounder-level equity — significant ownership in Europe’s autonomous pentesting leader.Modest salary (we build big, not pay big — yet).A great early-access platform and a clear mission: deliver consistent, better-than-human results.A hands-on, high-intensity environment, with 60-hour sprints when it counts.Direct, visible impact on the future of cybersecurity — your research becomes production reality.Join UsIf you want to build the next generation of autonomous security — where AI agents hack responsibly and continuously — we’d love to talk.Apply now and help us make our autonomous pentesting the best in class.