Senior Security Engineer (AI Safety), London, Lausanne

Your impact

As a Senior AI Security Engineer at Isomorphic Labs, you will operate at the absolute frontier of Artificial Intelligence and Cybersecurity. Reporting directly to the CISO, your mission is to architect, secure, and future-proof our pioneering AI-first platform and autonomous agentic workflows that drive the next generation of drug discovery pipelines.

In this high-impact role, you will design custom risk management frameworks tailored to the unique threat landscape of AI in life sciences, establish strict governance over critical model artifacts, and embed seamless security controls directly into distributed training and inference infrastructure. You will serve as the technical anchor bridging machine learning engineering, platform architecture, and adversarial defense.

Operating in a fast-moving, high-stakes environment, you will lead AI-specific incident response and pioneer model-driven defense mechanisms to automate threat hunting and mitigation. Your work will ensure that breakthrough scientific research is paired with an uncompromising, resilient defense safeguarding invaluable intellectual property while accelerating our innovation velocity.

What you will do

• Adversarial Threat Modeling: Conduct cutting-edge threat modeling focused on AI/ML vulnerabilities, engineering a comprehensive risk management framework tailored to an AI-powered life sciences platform.
• Model Artifacts Protection: Partner with AI teams to establish a granular inventory, classification and protection framework for ML models weights, code, and training data.
• Agentic AI & LLM Guardrails: Design and deploy scalable guardrails, robust sandboxing, and real-time monitoring controls for our extensive LLM ecosystem, including ADK, MCP, and autonomous agentic workflows.
• Collaborative Security Engineering: Partner with ML researchers and platform engineers to design and implement robust security controls across the entire ML lifecyclefrom data ingestion and training pipeline integrity to inference runtime controls.
• Model-Driven Incident Response: Act as the primary technical expert for AI security incidents, deploying advanced machine learning techniques to automate threat hunting, anomaly detection, and signatureless mitigation specific to LLMs.
• Compliance & Governance Automation: Bridge the gap between complex AI architectures and emerging global regulations (e.g., EU AI Act) by automating risk posture monitoring and continuous compliance metrics.
• AI Safety Frameworks: Pioneer the practical implementation of Isomorphic Labs' AI Safety standards, working in close coordination with Legal, Compliance, and external partners like Google DeepMind.

Skills and qualifications

Essential

• AI/ML Engineering: Deep conceptual and practical understanding of deep learning frameworks (e.g., JAX, PyTorch, TensorFlow), large-scale cloud training or inference infrastructure, and LLM ecosystems.
• Adversarial ML Mindset: Strong familiarity with AI security threat vectors (prompt injection, model inversion, data poisoning) and frameworks like OWASP Top 10 for LLMs or MITRE ATLAS.
• Agentic Frameworks Security: Proven experience securing agentic frameworks (ADK, MCP), agentic first infrastructure and implementing robust agent-to-agent (A2A) identity and access controls.
• Infrastructure Security: Solid proficiency in cloud platform security (GCP preferred), including container security, multi-cloud/SaaS integrations, and strict network isolation boundaries.
• Pragmatic Risk Assessment: Proven ability to define and execute pragmatic mitigation strategies that balance a rigorous security posture with the high-velocity needs of a world-class scientific research organization.
• Software Development & Automation: Ability to write production-grade code (Python preferred) to build custom security tooling, automate telemetry collection, or enforce policy layers.
• Translational Communication: Exceptional skills in navigating ambiguity, building trusted relationships with AI researchers, and clearly translating complex machine learning risks into actionable engineering tasks for leadership.

Nice to have

• AI Red Teaming: Experience conducting empirical vulnerability research or executing simulated attacks against LLMs, agent networks, and core ML backends.
• Regulated Environments: Prior exposure to BioTech, Pharma, and/or Deep Tech industries where data integrity, intellectual property protection, and regulatory compliance (GxP) are paramount.
• Education & Background: BSc, MSc, or PhD in Computer Science, Machine Learning, Cybersecurity, or a related quantitative field.
• Advanced Certifications: Relevant security or cloud credentials (e.g., OSCP, Professional Cloud Security Engineer) or specialized training in ML security.