Descripción de la oferta
ph3The Position /h3pAs a User Enablement Engineer within the Accelerated Compute Engineering (ACE) team, you will oversee the deployment, optimization, and day-to-day operation of user-facing software, application frameworks, and research environments across our High-Performance Computing (HPC) and AI Factory platforms. Your mission is to maximize the productivity of Roche's scientific community by providing frictionless access to cutting-edge computational tools. /ph3Job Responsibilities /h3h3User-Facing Platform Environment Management /h3pDeploy, support, and optimize user-facing environment portals and platforms, including Roche ROCs and SCOOP environments for HPC. Own the lifecycle and availability of standard software stacks, modules, and containerized application environments. Collaborate closely with infrastructure engineers to ensure user-facing portals interface seamlessly with core scheduling and data transfer layers. /ph3AI Factory Software Stack Productization /h3pArchitect and manage the deployment of the NVIDIA AI Enterprise software suite, ensuring proper integration and license optimization. Provide and support cutting-edge AI pipelines and development platforms, including NVIDIA NeMo for generative AI/LLMs and NVIDIA Omniverse for advanced industrial simulations and digital twins. Build, benchmark, and maintain highly optimized container builds for standard AI/ML frameworks (such as PyTorch and TensorFlow), incorporating deep learning inference servers like Dynamo-Triton to maximize GPU utilization. /ph3Scientific Consultation Technical Enablement /h3pAct as the technical entry point and consultant for research teams, assisting them in porting their code, algorithms, and models onto the ACE platforms. Develop, curate, and maintain developer documentation, example code repositories, and user guides to enable self-service onboarding. Diagnose and resolve complex software-layer bottlenecks, dependency conflicts, and compilation issues spanning scientific software packages, Python virtual environments, and GPU-accelerated libraries (e.g., CUDA, cuDNN). /ph3Qualifications /h3ullibEducation/Experience: /b Bachelor’s or an advanced degree in Computer Science, Data Science, Bioinformatics, Computational Chemistry, or a similar technical discipline; 5+ years of experience in an engineering role focused on scientific software compilation, user enablement, or DevOps for advanced research compute environments; Deep proficiency navigating and managing user applications within Enterprise Linux (RHEL/Ubuntu) multi-tenant systems. /lilibTechnical Business Skills: /b Experience configuring and optimizing NVIDIA AI Enterprise applications, specifically NeMo, Omniverse, and Triton Inference Server architectures; Deep understanding of modern AI/ML frameworks (PyTorch, Jax, TensorFlow) and their performance tuning on modern GPU architectures; Proficiency with environment management tools (such as Lmod/Environment Modules, Conda) and container technologies (Singularity/Apptainer, Docker); Strong scripting capabilities (Python, Bash) to automate user environment configurations and application testing pipelines; Enforce a strict configuration-as-code and infrastructure-as-code mindset, replacing manual interventions with repeatable automation scripts. /lilibLeadership Mindset: /b Highly focused on driving self-service capabilities and automating application delivery to scale support; Genuine passion for working with researchers and translating complex infrastructure into accessible services; Strong passion for innovation and staying current with the rapidly changing open-source and commercial AI/ML software landscape. /li /ulpRoche is an Equal Opportunity Employer. /p /p #J-18808-Ljbffr