Senior Security Engineer (AI Safety), London or Lausanne
job-boards.greenhouse.io
- Required language
- French conversational
- Job written in
- English
- Location
- Lausanne
- Work type
- Hybrid
Isomorphic Labs (IsoLabs) is a frontier‑AI company focused on accelerating drug discovery and ultimately solving disease. Building on the AlphaFold breakthrough, the firm creates predictive and generative AI models that design novel molecules and evaluate drug performance across many therapeutic areas. The Senior Security Engineer (AI Safety) will join an interdisciplinary team that blends machine learning, biology and cybersecurity to protect the company's AI‑first platform and autonomous‑agent workflows. In this role you will spend each day shaping the security architecture of IsoLabs' AI systems. Core duties include performing adversarial threat modelling for AI/ML vulnerabilities, establishing a detailed inventory and protection scheme for model artefacts such as weights, code and training data, and implementing guardrails, sandboxing and real‑time monitoring for large language model (LLM) ecosystems. You will work closely with ML researchers and platform engineers to embed security controls throughout the data‑ingestion, training and inference pipelines, and act as the primary technical lead for AI‑specific incident response, using advanced ML techniques to automate threat hunting and mitigation. Additionally, you will automate compliance monitoring for regulations like the EU AI Act and help operationalise IsoLabs' AI Safety standards together with legal and external partners. The position requires deep expertise in AI/ML engineering, including hands‑on experience with deep‑learning frameworks such as JAX, PyTorch or TensorFlow and large‑scale cloud training or inference environments. You must be familiar with AI security threat vectors, prompt injection, model inversion, data poisoning, and related frameworks such as OWASP Top 10 for LLMs or MITRE ATLAS. Proven experience securing agentic frameworks (ADK, MCP) and implementing robust agent‑to‑agent identity and access controls is essential, as is strong competence in cloud platform security, preferably on GCP, covering container security, multi‑cloud/SaaS integrations and network isolation. The role also calls for a pragmatic risk‑assessment mindset that balances rigorous security with the fast‑paced needs of scientific research. Nice‑to‑have skills include a track record of building compliance‑automation pipelines for emerging AI regulations, experience collaborating with external AI research groups such as Google DeepMind, and the ability to translate complex AI architectures into practical governance frameworks. The position is based in either London or Lausanne, offering the chance to work in a vibrant, collaborative environment alongside leading drug‑discovery scientists and machine‑learning specialists. What the role asks for: - Deep experience with JAX, PyTorch or TensorFlow - Proven expertise in AI security threat vectors - Experience securing agentic frameworks (ADK, MCP) - Strong cloud security skills, GCP preferred - Ability to design AI‑specific risk management frameworks - Experience with AI compliance (e.g., EU AI Act) - Familiarity with OWASP Top 10 for LLMs or MITRE ATLAS - Nice‑to‑have: collaboration with external AI research partners
Security Engineer: Lebenslauf-Vorlage
Bewirb dich mit einem Lebenslauf im Schweizer Aufbau — mit Beispieltext und den Anforderungen, die Security Engineer-Inserate am häufigsten nennen.
Lebenslauf-Vorlage Schweiz ansehenSecurity Engineer: was der Markt gerade verlangt
Security Engineer-Stellen gehören zu den regelmässig ausgeschriebenen Berufen auf SwissJobs.app.
Am häufigsten genannt in Security Engineer-Inseraten:
- Belastbar / flexibel8% der Inserate
- Teamfähig3% der Inserate
- Deutsch2% der Inserate
Weitersuchen
- Alle Cyber Security Jobs in der Schweiz
- Alle Jobs in Lausanne
- Lebenslauf als Security Engineer: Beispiel und Vorlage
- Lebenslauf mit KI erstellen
- ATS-Lebenslauf prüfen
- Bewerbungsschreiben für dieses Inserat