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
As a key driver of technological innovation, you will leverage cutting-edge AI technologies and advanced architectures to deliver strategic business value for Roche China. You will lead the end-to-end design, development, iterative enhancement, and ecosystem integration of the AI infrastructure, platform, and application adoption layers, ensuring all systems strictly adhere to enterprise security and compliance standards.
We are seeking a visionary architect who combines enterprise-level system design, platform engineering methodologies, and agile delivery with profound AI expertise. Your ultimate mission is to translate state-of-the-art AI capabilities into transformative, real-world healthcare solutions. In this role, you will be responsible for:
AI Platform Layer Design, Development, and Management
- Kubernetes Cluster Management: Lead the deployment, administration, and full lifecycle management of Kubernetes clusters, including node scaling, upgrades/rollbacks, certificate rotation, RBAC administration, and cluster teardowns. Develop robust automation code for cluster operations and drive federated cluster management initiatives.
- MLOps Platform Engineering: Architect, build, and manage MLOps platforms using open-source (e.g., Kubeflow, MLFlow) or commercial (Dataiku, Databricks) frameworks. Oversee the ML lifecycle by integrating core components (WebIDE, FeatureStore, ModelRegistry, AutoML, Pipelines, Model Serving) and implementing metrics for CMCT and Drift Detection.
- LLMOps Architecture: Design and develop platforms to support the Large Language Model (LLM) lifecycle. Integrate critical LLMOps components (Data Assets, Model Cards, SFT, OpenAI-Compatible Inference) and operationalize mainstream fine-tuning and inference frameworks (DeepSpeed, TorchTune, vLLM, SGLang).
- AI Gateway & Governance: Design and implement enterprise-level AI gateways and integration platforms. Establish governance frameworks (FinOps, Guardrails, MAAS PrivateLink) and integrate with domestic and international MAAS ecosystems (GCP Vertex AI, AWS Bedrock, Azure OpenAI, Ali DashScope).
- Familiarity with at least one major large-scale model training framework (e.g., Megatron-LM or MindSpore).
- Expertise in Ray for both distributed training and inference pipelines.
- Strong understanding of cross-node/multi-node scaling and distributed systems, essential for supporting next-generation Large Language Models (LLMs).
AI Application Layer Design, Development, and Management
- AI & Agentic Application Development: Architect, design, and develop highly scalable AI and Agentic applications, leveraging mainstream frameworks such as LangChain and LangGraph.
- System Integration: Integrate applications across advanced platform tools (Kubernetes, Slurm, MLOps, LLMOps, AI Gateway) and diverse data ecosystems (S3, RDBMS, VectorDB, GraphDB).
- Microservices Architecture & DDD: Drive Domain-Driven Design (DDD) to decompose APIs based on complex business requirements, defining clear service boundaries, contracts, and data models.
- Team Leadership & Delivery: Organize, lead, and mentor engineering teams to successfully deliver multi-format services, including APIs, WebUIs, SDKs, CLIs, and APPs.
- Engineering Excellence: Establish and enforce rigorous coding standards, comprehensive unit/integration testing protocols, and robust code review processes.
- Backend System Design: Design and optimize backend systems utilizing diverse API standards (REST, gRPC, GraphQL), database technologies (SQL, NoSQL), messaging systems (Kafka, RabbitMQ), and caching mechanisms.
- Architectural Strategy: Evaluate and implement appropriate architectural patterns (such as Event-Driven, CQRS, DDD, and Serverless) by conducting thorough trade-off analyses based on project needs.
- End-to-End Execution: Oversee the full lifecycle of AI software development, ensuring successful implementation and maintaining a track record of high-quality product delivery.
Platform Engineering, SRE, and Enterprise Integration
- Cloud-Native & Reusability Engineering: Drive platform hardening and component reusability engineering by leveraging mainstream CNCF projects (Prometheus, Grafana, OpenTelemetry, Ray, ELK, Kubeflow).
- Observability & FinOps: Design and integrate comprehensive observability solutions (Metrics, Logs, Tracing) and alerting systems to ensure transparent monitoring of application performance, usage, and FinOps costs.
- Non-Functional Requirements Assurance: Ensure all architectures strictly adhere to enterprise standards for high availability (HA), disaster recovery/rollback, security compliance, and component reusability.
- Enterprise Integration: Execute seamless integrations with the wider enterprise IT ecosystem, including SSO, LDAP, Git repositories, Artifact repositories, Cloud Landing Zones, Cloudflare, and AI Gateways.
AI Infrastructure Layer Design, Development, and Management
- Hardware & Storage Strategy: Evaluate, benchmark, and select AI infrastructure hardware (GPU and high-performance CPU bare metal servers), storage solutions (NAS, SAN), and coordinate physical data center requirements (power distribution, air/liquid cooling).
- Network Planning & Optimization: Architect network topologies and subnets, managing out-of-band networks, business networks, and high-performance RDMA passthrough networks (InfiniBand, RoCEv2). Deploy centralized and automated monitoring for all network equipment.
- OS & Toolchain Hardening: Administer Linux distributions and design automated security hardening protocols for large-scale server clusters. Manage supercomputing drivers and complete toolchains (Ansible, Packer, NVIDIA full-stack, RDMA full-stack, storage drivers, containerization).
- Cloud Infrastructure Automation: Manage enterprise-level private cloud environments, ensuring robust IAM, data center direct-connect, cloud service security, and data encryption. Drive Infrastructure as Code (IaC) automation utilizing Terraform, CloudFormation, and CLI tools.
- Privileged Access Management: Configure, manage, and maintain enterprise-level bastion hosts and oversee super-user access controls.
Demand Management, Process Design, and Agile Development
- Project & Delivery Management: Drive large-scale, role-based project execution applying standard Project Management Methodologies (PMM), ensuring alignment with pharmaceutical industry compliance standards.
- Agile Leadership: Champion Agile and Scrum methodologies. Perform WBS decomposition, manage project backlogs in Jira, and coordinate cross-team delivery through CI/CD pipelines (GitLab/GitHub, JFrog Artifactory).
- Governance & Process Design: Collaborate on the design and implementation of product-related operational processes (user training/certification, workspace/model/data asset lifecycle management, system disclaimers) to guarantee system safety and compliance.
Collaborations
- Project Documentation: Author and finalize critical architectural documents, including Solution Architectures, LLM specifications, and Installation Validation protocols.
- Knowledge Base Maintenance: Create, update, and govern comprehensive technical documentation across platforms like GitLab MD, Confluence, and GitLab Pages, including drafting detailed process design documents.
- Cross-Functional Team Leadership: Facilitate highly efficient collaboration across global and local cross-functional teams, ensuring seamless communication in both fluent Chinese and English.
- Mentorship: Providing architectural guidance and fostering team capability growth.
- Drive and lead efficient collaboration across multi-functions, global/local teams with fluent Chinese and English communication.
Qualifications
- 10+ years of experience as Solution Architect, with a minimum of 2+ years of hands-on experience specifically in AI Architecture or AI Platform design.
- Fluent bilingual capabilities in both English and Mandarin , enabling seamless, high-level architectural discussions with both global and local technical teams.
- Experience in the pharmaceutical, life sciences, or top-tier tech industries (with large-scale compute platform exposure) is highly preferred.
- Familiar with GxP relevant architecture design experience is highly preferred
- Deep mastery of the Kubernetes ecosystem and full-lifecycle cluster management.
- Extensive experience designing and building MLOps and LLMOps platforms (familiarity with Kubeflow, MLflow, Databricks, etc.).
- Deep knowledge of core ML components including Feature Stores, Model Asset Management, SFT (Supervised Fine-Tuning), and Model Serving.
- Familiarity with mainstream large model fine-tuning and inference acceleration frameworks (e.g., DeepSpeed, vLLM, SGLang, TorchTune etc.).
- High proficiency in RAG architecture design.
- Hands-on experience developing Agentic applications using mainstream frameworks such as LangChain and LangGraph.
- Familiarity with Vector Databases (VectorDB) and Graph Databases (GraphDB).
- Rich experience in High Availability (HA) system design, Disaster Recovery, RBAC, and integrating enterprise-level components (e.g., SSO, LDAP, Bastion hosts).
- Proven ability to ensure that architectural designs strictly adhere to security compliance standards and FinOps governance requirements.
- In-depth understanding of Cloud-Native architectures (e.g., AWS, Alibaba Cloud).
- High proficiency in the CNCF mainstream open-source technology stack, with particular expertise in observability tools (e.g., Prometheus, Grafana, OpenTelemetry, ELK).
- A strong platform engineering mindset with extensive hands-on experience using Terraform, Ansible, Packer, and CloudFormation for automated deployment and system security hardening.
- Mastery of at least one core backend development language (e.g.Python, Go).
- Expertise in building robust, end-to-end pipelines utilizing tools such as GitLab CI/CD and JFrog.
- Strong familiarity with foundational AI hardware (GPU/high-performance CPU) and storage technologies (NAS/SAN).
- Proven ability to conduct performance benchmarking and hardware selection for high-compute clusters.
- Profound understanding of data center network architectures, with expertise in planning and configuring out-of-band networks and high-performance RDMA passthrough environments.
- Mastery of Linux (RHEL/CentOS) operating systems and high proficiency in configuring the NVIDIA driver full-stack and supercomputing driver toolchains.
- High proficiency in Scrum/Agile development processes and necessary tools.
- Demonstrated track record of leading large-scale AI platforms or projects and mentoring cross-functional teams to successful delivery.
Who we are
A healthier future drives us to innovate. Together, more than 100’000 employees across the globe are dedicated to advance science, ensuring everyone has access to healthcare today and for generations to come. Our efforts result in more than 26 million people treated with our medicines and over 30 billion tests conducted using our Diagnostics products. We empower each other to explore new possibilities, foster creativity, and keep our ambitions high, so we can deliver life-changing healthcare solutions that make a global impact.
Let’s build a healthier future, together.
Roche is an Equal Opportunity Employer.
