Lead AI Engineer (Agentic Systems)

About the role

Roche is hiring a Lead AI Engineer focused on agentic systems for its Diagnostics Commercial business. The role defines the global blueprint for multi-agent systems that orchestrate complex workflows, connect with legacy enterprise data and make autonomous decisions safely. It is a senior, high-impact position combining technical leadership with strategic influence.

Key responsibilities

• Act as chief architect for the Commercial GenAI platform strategy, designing multi-agent orchestration blueprints, state-management protocols and shared tool-use boundaries.
• Lead the evaluation and adoption of agentic paradigms such as hierarchical multi-agent teams, reflection loops, knowledge semantic models and orchestration frameworks like LangGraph, AutoGen or custom orchestrators.
• Establish enterprise standards for LLMOps and Agent-Ops, including benchmarking, fallback mechanics, semantic caching, automated cost containment and guardrails against runaway autonomy.
• Work with senior executives, product heads and legal/compliance councils to turn large business problems into compliant, agent-driven software.
• Mentor senior developers and AI engineers across global digital hubs through matrixed leadership and a community of practice.

Required profile

• Postgraduate degree (Master's or PhD preferred) in Computer Science, Software Engineering, AI/Machine Learning or an equivalent field.
• 8+ years of hands-on software engineering, including at least 3+ years architecting, deploying and scaling complex Generative AI and LLM workflows in production.
• Experience designing automated systems in highly regulated industries (healthcare, finance or aerospace) is highly desirable.
• Ability to influence without formal authority and to present technical roadmaps to C-suite stakeholders.

Required skills

• State management in agentic networks, including DAGs for LLM reasoning, agent negotiation loops and persistent cognitive memory.
• Microservices architecture, secure API orchestration, semantic routing and cloud-native scaling (AWS/Azure/GCP).
• Advanced vector search, graph databases (e.g. Neo4j) and hybrid retrieval.
• Custom evaluation frameworks for non-deterministic AI systems.
• Knowledge of AI safety, algorithmic bias mitigation and data privacy compliance.

What we offer

• Executive-level total rewards package with competitive base pay and performance bonuses.
• Comprehensive health and wellness benefits.
• Flexible executive hybrid work arrangements.