We are seeking a Semantic Architecture & Context Engineering Lead to help define and scale the semantic foundation underpinning our enterprise data, analytics, AI, and agentic ecosystem. As organizations increasingly consume data through natural-language interfaces, generative AI, intelligent agents, predictive models, and automated workflows, the quality of the underlying context becomes as important as the quality of the underlying data. Business concepts, metrics, relationships, definitions, metadata, instructions, retrieval strategies, and decision rules must be structured in ways that are consistent, machine-readable, reusable, governed, and optimized for AI consumption. This role will establish the practices, standards, and reusable patterns through which business and data context is represented across our data products, analytical models, GenAI solutions, and agent architectures.
The individual will work horizontally across Data, Analytics, AI, Product, Engineering, Architecture, and business domain teams. They will partner closely with functional ownersincluding data product managers, AI engineers, data scientists, architects, and business subject-matter expertsto ensure semantic and contextual structures are designed consistently while remaining appropriate to individual domains and use cases. Critically, this is not solely a governance or documentation role. The Semantic Architecture & Context Engineering Lead will establish an experimental discipline around context, using structured evaluation, A/B testing, production telemetry, and AI-assisted techniques to continuously improve the accuracy, precision, reliability, performance, and usability of AI-enabled solutions.

Job Responsibilities

• Enterprise Semantic & Context Architecture.Define the enterprise framework for representing business meaning and context across data, analytics, AI, and agentic solutions. Establish standards and reusable patterns for semantic models, business concepts, metrics, entities, relationships, taxonomies, metadata, definitions, contextual instructions, and knowledge structures. Develop a common approach for translating human business concepts into machine-readable and AI-consumable representations. Define principles for semantic interoperability across data products and domains, minimizing conflicting definitions, redundant logic, and fragmented representations of common enterprise concepts. Establish semantic and context architecture patterns that can be applied across structured, semi-structured, and unstructured information.
• AI & Agent Context Engineering -Define best practices for structuring the context provided to GenAI and agentic solutions, including system instructions, business rules, semantic metadata, examples, retrieval context, tool descriptions, entity relationships, and domain knowledge. Partner with AI engineering and architecture teams to design context patterns for RAG, natural-language-to-data, reasoning, agent tool use, memory, workflow orchestration, and agent-to-agent interaction. Develop reusable context architectures for solutions such as Databricks Genie spaces/rooms, enterprise AI assistants, analytical agents, and domain-specific agents. Ensure context structures clearly establish authoritative definitions, permitted sources, relationships, hierarchies, temporal logic, calculation rules, and other constraints required for reliable AI reasoning. Help establish standards for context isolation, inheritance, reuse, versioning, and lifecycle management across an expanding portfolio of AI solutions.
• Data Product Semantics -Partner with data product managers and domain teams to strengthen the semantic layer of enterprise data products. Establish standards for business definitions, logical models, physical-to-logical mappings, metrics, dimensions, entity relationships, metadata, lineage, data-quality expectations, and consumption guidance. Ensure data products contain sufficient semantic context for use by human analysts, BI platforms, machine-learning models, GenAI applications, and autonomous agents. Promote reusable semantic contracts that allow downstream solutions to interpret data consistently without recreating business logic independently. Identify common concepts and relationships that should be standardized across multiple data domains while preserving appropriate domain ownership.
• Context Experimentation & Evaluation -Establish a systematic testing framework for semantic and contextual design. Design and lead experiments to evaluate alternative context structures, instructions, examples, semantic representations, retrieval strategies, and metadata configurations.
  Use A/B testing, offline evaluation, golden question sets, adversarial testing, production telemetry, and other methods to measure changes in: - answer accuracy, precision and recall, grounding and source fidelity etc
• Semantic Lifecycle & Governance - Establish lifecycle practices for creating, reviewing, approving, versioning, testing, deploying, monitoring, and retiring semantic and contextual assets. Define ownership models distinguishing enterprise standards from domain-level business ownership. Ensure semantic assets can be traced to authoritative business definitions and trusted data sources.
• AI-Enabled Semantic Engineering - Identify opportunities to embed AI throughout the semantic lifecycle. Use AI-assisted techniques to accelerate metadata generation, concept extraction, taxonomy development, relationship identification, mapping, documentation, semantic validation, and quality assurance.

• Partnership & Adoption - Act as a subject-matter expert and internal advisor on semantic architecture and context engineering. Partner with Data Product, Data Engineering, AI/ML, Business Intelligence, Product, Design, Architecture, and business teams to embed semantic best practices into delivery.
  Conduct design reviews and working sessions for priority data and AI solutions. Develop playbooks, templates, education, and reusable implementation patterns that allow teams to adopt semantic practices without requiring centralized execution.
  Build a community of practice across data and AI teams to continuously advance enterprise capabilities. Translate highly technical semantic concepts into clear implications for product owners, executives, and business stakeholders.

Required Qualifications

• Significant experience in one or more of the following areas: semantic architecture, data architecture, knowledge engineering, ontology development, data product management, AI/ML engineering, information architecture, or advanced analytics.
• Strong understanding of enterprise data architecture, including logical and physical data models, metadata, data products, metrics, dimensions, lineage, and data governance.
• Demonstrated understanding of modern GenAI architectures, including LLMs, RAG, vector retrieval, natural-language-to-data solutions, prompt/context engineering, and agentic architectures.
• Experience translating complex business concepts and domain knowledge into structured technical representations.
• Strong analytical orientation with experience designing experiments, evaluations, or other quantitative methods for assessing solution performance.
• Ability to operate effectively across business, product, data, engineering, and architecture organizations.
• Strong written and verbal communication skills and the ability to influence without direct ownership of participating teams. Ability to balance enterprise standardization with pragmatic delivery and domain-specific requirements.
  

Preferred Qualifications

• Experience with Databricks, including Unity Catalog, Delta Lake, Databricks SQL, MLflow, Vector Search, and/or Genie.
• Experience with enterprise semantic layers, metric stores, business glossaries, metadata platforms, data catalogs, or ontology-management technologies.
• Familiarity with knowledge graphs and semantic technologies such as RDF, OWL, SKOS, SHACL, SPARQL, Neo4j, or similar technologies.
• Experience with RAG, GraphRAG, embedding models, vector databases, semantic search, and knowledge-grounded AI systems.
• Familiarity with AI evaluation frameworks, including golden datasets, automated evaluators, LLM-as-judge techniques, model observability, or production AI monitoring.
• Experience operating in a highly governed or regulated enterprise environment.
• Financial services and/or Human Resources domain experience is advantageous but not required.