About this role
Amgen is seeking a hands-on Machine Learning Engineer (Level 4) to design, build, and operationalize generative and agentic AI systems that are safe, auditable, and production-ready. The role involves owning LLM and agent capabilities, evaluation frameworks, and guardrails, while collaborating with data science, product, platform, and compliance teams to transition models from prototypes into trusted enterprise services. Responsibilities include designing and deploying LLM solutions and agent frameworks, building evaluation pipelines (automated metrics, unit/integration tests, adversarial/red-team testing, human-in-the-loop), and creating guardrails for prompt patterns, policy enforcement, and output validation. You will also prototype agentic behaviors such as RAG, multi-step planners, and tool-use interfaces, while implementing MLOps best practices including model versioning, CI/CD, and scalable serving. Required qualifications include 5 to 9 years of applied machine learning or software engineering experience, with at least 2 years specifically on production ML systems, and a B.Tech, M.Tech, or MS in Computer Science. Candidates must possess strong hands-on experience with LLMs (fine-tuning, instruction tuning, retrieval augmentation, embeddings), agent frameworks, and solid software engineering skills in Python and deep learning frameworks like PyTorch, JAX, or TensorFlow. Experience with cloud deployment, Kubernetes, and model serving is required. Preferred qualifications include prior work on multi-agent systems, benchmarking, background in regulated industries like pharma or healthcare, and contributions to open-source or academic work in LLMs. The role requires a security and compliance mindset, pragmatic execution, and the ability to engage cross-functional stakeholders.