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…