YOUR CHALLENGE:

• Own and evolve enterprise agent frameworks and reference architectures (orchestration, tool/function calling, workflow integration, memory/state and retrieval/grounding) with clear boundaries, escalation rules and hand‑off guidance
• Build and maintain an "Agent Factory” of reusable components and blueprints (agent templates, tool adapters, prompt modules, evaluation harnesses and golden conversations) to accelerate Domain Pod delivery and maximise reuse across Sunrise
• Define and implement agent safety layers, guardrails, policy enforcement, constrained tool access, bounded autonomy and human‑in‑the‑loop approval/override patterns, and contribute to red‑teaming, abuse testing and robustness testing
• Define prompt standards and libraries; engineer prompts and system instructions for grounding, tool use, structured outputs and safe behaviour; enforce strict versioning, peer review, regression testing and change control
• Design multilingual (DE/FR/IT/EN) customer and operational conversation flows with tone/style guidance, recovery paths and escalation patterns aligned to company guidelines and regulatory requirements
• Partner with engineering teams to integrate agents into enterprise channels and workflows by defining tool schemas, authorisation checks, error handling, state management and full end‑to‑end observability (traces, logs and evaluation telemetry)
• Define release readiness criteria (quality thresholds, safety checks, monitoring coverage, rollback/kill switch, runbooks) and support Pods through go‑live, hypercare and post‑implementation reviews
• Monitor and continuously improve production behaviour (quality, grounding/hallucination risk, task success, escalation/containment and cost/latency), support incidents/postmortems and coach embedded specialists for systematic iteration

YOUR SKILLS:

• Bachelor’s/Master’s degree in HCI, Engineering, Computer Science, Cognitive Science, Linguistics, Data/AI, Interaction Design or equivalent practical experience
• Hands‑on experience building and shipping conversational AI and agentic workflows (tool calling, grounding/RAG, multi‑step workflows, structured outputs)
• Strong prompt engineering and conversation design skills, including multilingual experience (DE/FR/IT/EN) and UX‑oriented dialogue design
• Practical experience with systematic evaluation (rubrics, golden sets), prompt/agent regression testing and disciplined versioning/change control
• Working knowledge of modern LLM patterns (function/tool calling, structured outputs, retrieval/grounding, context management and reliability techniques)
• Experience designing safety and privacy controls (guardrails, bounded autonomy, HITL, PII handling, policy compliance) and collaborating with governance, security, privacy and risk functions
• Experience partnering in cross‑functional product teams and matrixed Hub‑and‑Spoke delivery environments; ability to coach others through reusable patterns
• Experience with production operations for AI solutions: observability/telemetry, monitoring/alerting, incident response and continuous quality improvement
• Business‑fluent English required; German is an advantage depending on stakeholder scope

At Sunrise, you’ll be working in a dynamic and multinational environment where every voice is heard, perspectives are shared and values are lived. Diversity isn’t just a program for us, it’s part of our DNA. By celebrating our differences, we achieve great things together.

Sunrise has partnered with «Advance – Gender Equality in Business». We’ve also been awarded the Swiss LGBTI label and «Fair-ON-Pay Advanced» certification as recognition for offering all our employees equal pay for work of equal value.

Ready to rise to the challenge? Then join the team and actively shape our future!

#ChallengersWanted

Information for agencies: Application dossiers sent or uploaded by placement agencies or similar are not desired, will therefore not be considered and will be deleted.

#LI-SJ1