Machine Learning Engineer, AI Studio

Amgen is a biotechnology company focused on developing and delivering innovative medicines for patients with serious illnesses. The Machine Learning Engineer will independently own production components within enterprise AI products and automation solutions, designing, releasing, diagnosing, and supporting components across machine learning, Generative AI, data, retrieval, agents, and workflow systems. The role also involves applying responsible AI, security, privacy, validation, monitoring, and lifecycle operations practices while guiding associate engineers.

Responsibilities:

• Define component boundaries, intended use, acceptance criteria, non-functional requirements, decision consequences, support expectations and technical estimates with product and architecture partners
• Design and implement maintainable Python, SQL, API, data, model, retrieval, agent-tool and workflow components with clear contracts, configuration, testing, error handling and documentation
• Apply EDA, feature engineering, supervised or unsupervised methods, baselines, cross-validation, leakage prevention, calibration, subgroup, threshold, explainability and error analysis where relevant
• Build GenAI, NLP, RAG and bounded agent components using structured output, embeddings, hybrid search, reranking, provenance, citations, permissions, approvals, retries and recoverable failure behaviour
• Engineer batch or event-driven data, document, feature, embedding, label and evaluation pipelines with schema validation, lineage, provenance, access control and consistency checks
• Define representative evaluation for model quality, uncertainty, retrieval, grounding, citations, task success, tool correctness, safety, latency, cost and user impact
• Release and support components using cloud services, containers, CI/CD, versioning, monitoring, rollback, incident response and runbooks; lead diagnosis of moderately complex failures
• Apply security, privacy, Responsible AI, validation, auditability, human oversight and applicable GxP controls; contribute reusable assets and guide Associate engineers on familiar work

Requirements:

• Master's degree
• Bachelor's degree and 2 years of Computer Science, IT or related field experience
• Associate's degree and 6 years of Computer Science, IT or related field experience
• High school diploma / GED and 8 years of Computer Science, IT or related field experience
• Demonstrated ownership of at least one production software, data, ML, GenAI or automation component
• Strong hands-on proficiency in Python and SQL, with sound software-engineering and testing practices
• Strong capability in at least one of classical ML, GenAI/RAG/agents or MLOps/platform engineering, with working knowledge of adjacent areas
• Advanced ML and deep learning: Experience with PyTorch, TensorFlow, Hugging Face, scikit-learn, XGBoost, PyMC, computer vision, NLP, GNNs, causal inference or uncertainty estimation
• Advanced GenAI and knowledge systems: Experience with LangChain, LangGraph, LlamaIndex, Semantic Kernel, AutoGen, CrewAI, hybrid retrieval, knowledge graphs, graph RAG or evidence verification
• Cloud, data and MLOps: Experience with AWS, Bedrock or SageMaker, Databricks, Spark, Kubernetes, infrastructure as code, MLflow, Airflow, Kubeflow or GitHub Actions
• Agents and human-AI workflows: Familiarity with MCP-style integration, agent tracing, adversarial testing, durable workflows, permissions, human review, BI or process automation
• Regulated enterprise delivery: Experience in healthcare, life sciences, GxP, validated systems or another regulated or high-impact environment
• Independent problem solving and sound component-level technical judgment
• Clear communication of assumptions, evidence, trade-offs, risks and support implications
• Strong collaboration with business SMEs, product, architecture, software, data, platform, evaluation and control partners
• Ownership, reliability and disciplined follow-through from design through production support
• Ability to guide junior engineers and learn new tools through evidence-based experimentation