Senior AI Platform Engineer - Memory & Graphs

Position

Senior‑Most Technical Knowledge & Context Engineer  will design and operationalize a centralized multimodal memory architecture spanning vector retrieval, knowledge graphs, ontologies, metadata systems, runtime memory injection, and lineage and governance frameworks.

This role sits at the intersection of AI engineering, distributed systems, and applied intelligence. You will lead the strategy and execution for how enterprise knowledge is structured, represented, governed, retrieved, and operationalized across agents, copilots, automation systems, and customer‑facing AI products.

Importantly, you will sit at the cutting edge of building the cognitive substrate for an AI-native enterprise. Come join us as a hands‑on thought leader who innovates by doing.

In addition to the specific responsibilities outlined below, employees may be required to perform other duties as assigned by the Company.

• Build foundational content intelligence systems:
  Design systems for ingestion, indexing, embeddings, metadata, retrieval, lineage, governance, and auditability that can support both internal and customer‑facing AI use cases. This includes customers, media and marketing assets, product features, production code, websites, game sounds, customer service experiences, operational workflows, and other enterprise content.
• Establish enterprise knowledge graphs and ontologies:
  Define and implement a regulatory‑compliant knowledge graph strategy that creates deep context about Fan Duel’s products, employees, operations, customers, and systems. Own the design of graph databases, semantic models, ontologies, entity‑resolution patterns, and relationships across vector and non‑vector data. Build the connective tissue that allows AI systems and agents to reason over enterprise context with precision, transparency, and control.
• Design secure memory patterns for agents:
  Create reusable design patterns for how AI agents acquire, store, retrieve, update, and discard context during runtime. This includes short‑term memory, long‑term memory, episodic memory, summarization, context injection, retrieval augmentation, and governed memory sharing across tools and systems. Ensure memory systems are efficient, secure, auditable, and appropriate for regulated environments.
• Champion responsible AI development:
  Ensure knowledge, retrieval, graph, and memory systems meet regulatory requirements, ethical standards, privacy obligations, and responsible gaming principles. Build safeguards, access controls, provenance, explainability, monitoring, and auditability into the platform from day one.

• Reusable runtime memory patterns that are being used by multiple agents to securely acquire, retrieve, summarize, and apply context during execution.
• A production‑grade graph and ontology framework connecting products, customers, employees, operations, systems, and content with clear lineage, access controls, and regulatory compliance.
• The first context sets of centralized, governed multimodal vector store and retrieval layer supporting AI applications across customer, product, marketing, engineering, operations, and support domains.
• Teams can build AI solutions faster because retrieval, memory, metadata, governance, and knowledge graph capabilities are available as shared primitives rather than bespoke pipelines.

• 7+ years of engineering experience, preferably working with large distributed systems that support a mix of data and software development activities.
• Track record of taking products from concept to launch in fast‑moving, ambiguous environments.
• Practical fluency with generative AI tools and concepts—especially graph, RAG, Agentic
  
  RAG, fine‑tuning, and anti‑RAG patterns.
• Experience building or operating vector and graph DBs, ontology, semantic search, and runtime memory evaluations.
• Ability to operate autonomously, create clarity from ambiguity, and influence across a matrixed organization.
• Strong communication skills—able to translate AI concepts into clear value narratives for both technical and non‑technical stakeholders.
• High ownership mindset with a bias for action and outcomes over…