AI Engineer

Roche; Basel; On-site; CHF 140,000 - 210,000; Full time

Job summary

Roche in Basel seeks an AI Engineer Level 6 to lead end-to-end AI delivery for Pharma Technical Operations. You will design architectures, build data pipelines, and implement GenAI solutions with a focus on reliability, compliance, and measurable business value.

You will mentor engineers, own solution delivery, and collaborate with business, IT, and quality teams to translate operational challenges into robust AI-enabled products across production and analytics.

Qualifications

• Senior AI engineer with hands-on delivery ownership.
• Experience with GenAI, LLMs, and RAG architectures.
• Ability to design scalable AI architectures and data pipelines.

Responsibilities

• Own end-to-end AI delivery from problem understanding to production.
• Design scalable AI application architectures (data flows, API, UI, deployment).
• Develop GenAI solutions, agent workflows, and evaluation approaches.
• Ensure engineering quality through testing, CI/CD, observability, and documentation.
• Maintain production reliability, security, cost-awareness, and continuous improvement.
• Translate business needs into technical solutions with measurable impact.

Skills

AI engineering depth

Software engineering

Cloud & DevOps

Data engineering

Evaluation methods

Job description

At Roche you can show up as yourself, embraced for the unique qualities you bring. Our culture encourages personal expression, open dialogue, and genuine connections, where you are valued, accepted and respected for who you are, allowing you to thrive both personally and professionally. This is how we aim to prevent, stop and cure diseases and ensure everyone has access to healthcare today and for generations to come. Join Roche, where every voice matters.

The Position

At Roche, we are committed to delivering greater benefits to our patients by applying digital, data, and AI capabilities to real-world operational challenges. The AI Engineer Level 6 is a senior hands-on technical contributor within the AI Engineering Cluster, responsible for designing, delivering, evaluating, and operating robust AI solutions across Pharma Technical Operations.

This role owns meaningful parts of end-to-end AI delivery, including solution architecture, data pipelines, machine learning and GenAI systems, agentic workflows, AI evaluation harnesses, backend services, user interfaces, and production operations. The role partners with business, IT, quality, and technical experts to translate complex operational challenges into reliable, scalable, compliant, and measurable AI-enabled products.

Role description

The role is a senior AI engineering role for someone who combines strong technical depth with practical delivery ownership. The AI Engineer Level 6 is expected to independently lead solution delivery, make sound architectural decisions, mentor other engineers, and ensure AI solutions create measurable business value while meeting reliability, scalability, evaluation, agent management, and compliance expectations.

AI Solution Leadership & Delivery

• End-to-end delivery ownership: Own the technical delivery of AI solutions from problem understanding and design through build, deployment, monitoring, evaluation, adoption, and continuous improvement.
• Solution architecture: Design scalable AI application architectures, including data flows, model integration, GenAI orchestration, APIs, user interfaces, agent workflows, and production infrastructure.
• GenAI and agentic AI: Design and implement advanced GenAI solutions, including RAG architectures, multi-step workflows, tool-using agents, LLM orchestration, and evaluation approaches.
• Engineering quality: Apply and help evolve strong engineering practices, including automated testing, code reviews, CI/CD, observability, documentation, and maintainability.
• Production performance: Ensure AI solutions are reliable, secure, monitored, cost-aware, and continuously improved in production or production-like environments.
• Value delivery: Translate business needs into technical solutions and help define success metrics that demonstrate operational and business impact.

Modern AI Engineering Capabilities

• Evaluation framework design: Design and implement data science-driven evaluation frameworks for AI, GenAI, RAG, and agentic AI systems, including offline benchmarks, golden datasets, synthetic test cases, human evaluation workflows, and automated regression testing.
• AI harness development: Own the development of AI engineering harnesses for repeatable experimentation, evaluation, validation, and comparison of models, prompts, retrieval strategies, agent behaviors, and tool integrations.
• Agentic AI solution ownership: Design and deliver production-grade agentic AI systems, including agent orchestration, tool integration, task decomposition, state handling, memory patterns, guardrails, and human escalation paths.
• Agent lifecycle management: Manage deployed agents across their lifecycle, including versioning, testing, release management, monitoring, incident response, continuous improvement, and retirement.
• Performance and risk monitoring: Implement and operate observability for AI systems, including quality metrics, business impact metrics, cost, latency, failure modes, drift, hallucination risk, grounding quality, and operational reliability.
• Model and workflow comparison: Lead structured experiments to compare foundation models, prompts, embeddings, retrieval approaches, orchestration patterns, and agent configurations.

Technical Expertise

• AI engineering depth: Apply strong knowledge of machine learning, predictive modeling, time-series, computer vision and multimodal AI, as well as LLMs, RAG, agentic AI, prompt engineering, model evaluation, and AI application design.
• Software and platform engineering: Build robust Python-based services, APIs, microservices, dashboards, and integrations using modern software engineering practices.
• Cloud and DevOps: Design and deploy solutions using AWS, Docker, CI/CD pipelines, monitoring, and cloud or hybrid deployment patterns.
• Data engineering: Work with SQL, operational data, time-series data, structured and unstructured data, and data pipelines required for AI products.
• Evaluation methods: Use quantitative metrics, qualitative review, human-in-the-