ML Software Engineer - AI Member Systems; AIMS

Opportunity

The AI for Member Systems (AIMS) organization sits at the core of this experience, building and operating the AI systems that power recommendations, personalization, search, discovery, and messaging. AIMS is developing a foundational new architecture, building a shared intelligence layer that unifies how personalization works across every Netflix surface. This layer orchestrates foundation models, retrieval, ranking, and policy through a governed runtime, replacing fragmented surface‑by‑surface systems with a coherent platform that compounds learning across the entire member experience.

We are looking for a senior ML Software Engineer (L6) to work at the intersection of this intelligence layer and the platform infrastructure that supports it. This role focuses on the systems that connect AIMS' AI capabilities to the broader Netflix ML and serving ecosystem, ensuring they are scalable, well‑integrated, and built for the next generation of AI‑powered member experiences.

• Design and build the platform integration layer between AIMS' intelligence systems and Netflix's ML infrastructure, serving layer, and product engineering teams.
• Drive architectural decisions on how AIMS capabilities (ranking, retrieval, orchestration, foundation models) connect to shared platform services including model serving, inference infrastructure, feature stores, and experiment platforms.
• Own the technical design of key integration points between the intelligence layer and the member experience serving stack, ensuring clean contracts, reliable handoffs, and operational excellence.
• Architect “paved paths” for AIMS application teams to build and deploy ML models and GenAI capabilities through consistent patterns in data access, training, evaluation, deployment, and monitoring.
• Design reusable, horizontal infrastructure components that multiple AIMS teams can adopt, avoiding duplication and ensuring improvements propagate across surfaces.
• Scope and de‑risk new architectural directions through prototypes and proofs of concept, especially where optimal abstractions between the intelligence layer and platform infrastructure are not yet clear.
• Shape requirements with platform and infrastructure partners, translating AIMS' needs into actionable capabilities and ensuring architectural alignment with broader Netflix platforms.
• Drive technical excellence across the systems you touch, raising the bar on reliability, observability, performance, and cost‑effectiveness.
• Partner closely with ML practitioners (research scientists, research engineers) to ensure platform and infrastructure decisions support rapid model development and experimentation.
• Mentor and elevate engineers across AIMS through technical guidance, design reviews, and championing principled engineering practices.

• Significant experience designing, building, and operating production ML systems end‑to‑end, spanning data access and pipelines, training and evaluation workflows, model serving, and online inference for high‑traffic products.
• Strong software engineering fundamentals with deep Python expertise, plus working proficiency in a JVM language (Scala or Java) to build and integrate with large‑scale platform services.
• Experience building platform and integration layers that connect ML capabilities to shared infrastructure (serving, inference, feature/embedding stores, experimentation), with clean APIs/contracts and reliable operational handoffs.
• Demonstrated ability to work at the boundary between ML teams and platform/infrastructure teams, translating requirements in both directions and driving alignment.
• Fluency with modern ML and GenAI patterns, including recommendation/ranking systems, LLM serving, and agentic architectures.
• Proven ability to identify common patterns and design frameworks and abstractions that are flexible, extensible, and easy for engineers to adopt.
• Hands‑on ability to scope and validate architectures through prototypes, turning ambiguous problems into concrete proposals and reference implementations.
• Comfortable influencing technical direction across multiple teams without formal authority, building consensus…