First seen by Alion on Nov 18, 2026.
This a Full Remote job, the offer is available from: United States, California (USA), New Jersey (USA), New York (USA), Texas (USA)
Who We Are:
TKO Group Holdings, Inc. (NYSE: TKO) is a premium sports and entertainment company. TKO owns iconic properties including UFC, the world’s premier mixed martial arts organization; WWE, the global leader in sports entertainment; and PBR, the world’s premier bull riding organization. Together, these properties reach 1 billion households across 210 countries and territories and organize more than 500 live events year-round, attracting more than three million fans. TKO also services and partners with major sports rights holders through IMG, an industry-leading global sports marketing agency; and On Location, a global leader in premium experiential hospitality.Why this role exists
TKO is consolidating the data estates of five businesses onto one enterprise data platform.
The Principal Data Architect & Modeler is going to lead data architecture role at TKO. You define the standards, the data model, the framework and the roadmap for the platform. Product, Engineering and Data Science build to your recommendations; Data Governance and Master Data Management set the policies your designs make enforceable.
This role is hands on as well as an influencer. You will design & write models, semantic definitions, standards and architecture decision records.
The north star
The platform exists to deliver three enterprise views, on one AI-ready foundation, built iteratively.
Fan 360 - one resolved identity for the consumer across UFC, WWE and PBR: tickets, viewership, merchandise, membership, engagement and consent, with the lineage that makes cross-brand use defensible.
Event 360 - every live event, broadcast and activation on one spine: venue, gate, attendance, rights, sponsorship, hospitality and media performance, at the grain each business needs.
Customer 360 - the B2B side of TKO: rights holders, sponsors, partners, hospitality buyers and agency clients across IMG and On Location, resolved to golden records and linked to the events and fans they touch.
Employee 360 - Trusted employee domain aligned to enterprise HR systems and P+C priorities; sequenced behind higher-value domains
AI-readiness is central theme, not a feature appended at the end. Agents and reporting needs to be based on a common definition of metric and attributes, context and data relationships embedded within the semantic layer access to information based on user roles..
None of this is built in one pass. The job is to set the end state, then sequence the increments - one entity, one conformed dimension, one certified metric set at a time - so each release puts a usable slice of a 360 into production and the model improves under use.
What you will own
The enterprise data framework and roadmap. Define the target-state architecture and the published roadmap from today's estate to the north star. Set the reference patterns for how data is ingested (batch, CDC, streaming, API), how it is curated through the medallion from landing to marts, how it is served to BI, agents and applications, and how data science and ML consume it (feature-ready datasets, training and retrieval pipelines, in-platform serving). Define the environments and promotion path from development to production, and the standards each layer must meet to pass. Engineering builds the pipelines, Product owns the data products, Data Science builds the models; your framework is what they all build against.
The canonical enterprise model. Define the dimensional and conformed-dimension standards that govern how raw data becomes curated data: grain, conformance, slowly changing dimension handling, surrogate and business keys, and the shared dimensions that must mean the same thing in every brand. Design for modularity, so that the fifth domain costs less than the first and new properties, partnerships and source systems are absorbed rather than re-engineered.
Master data design. Design the master data architecture across fan, event, rights holder, sponsor and partner, employee and vendor, in partnership with Master Data Management and Governance, which owns the policy, stewardship and governance framework your designs implement. Choose and justify the implementation style per domain. Design entity resolution as a cascade - deterministic, probabilistic, machine-learned, human-stewarded - with attribute-level survivorship, versioned match rules, and merges that are defensible and reversible. Treat consent state, opt-out signals and retention obligations as master data carried with identity across brands.
The semantic layer and its hierarchy. Own the enterprise-certified definitions for the concepts the whole company shares - fan, event, attendance, revenue, engagement, rights - and the domain models that inherit from and extend them. Set the metric tiers, the certification bar at each tier, and the promotion path from exploratory to certified. Decide where each definition physically lives across Snowflake semantic views, dbt, the BI semantic layer and the Open Semantic Interchange specification, and keep those representations in sync. TKO already runs conversational analytics agents in production for business users; the accuracy of every one of them is bounded by this model.
Governance made enforceable in the platform. You do not own governance. You design the architecture that makes it real: how the catalog is populated and kept current, how lineage is captured from source through transformation to semantic layer to agent, where data quality rules execute and how results surface, what observability every data product carries by default, and how classification and access boundaries travel with the data into every tool and agent. Convert standards into automated checks - model contracts, linting, pre-merge data diffs, quality thresholds - so governance happens by rule rather than by inspection.
Architecture decisions and the paved road. Establish the architecture decision record practice, stored with the code. Run an advisory forum rather than an approval board. Publish and maintain the reference implementations, templates and modeling patterns that make the standard the path of least resistance. Define the target model for each domain before its migration wave, and the parity criteria that decide whether a cutover is complete.
How you will work
The role reports to the VP, Data Engineering & Science within TKO's Enterprise Data organization and works across four partnerships, each with a deliberate boundary.
Data Governance & MDM: Partner owns policy, stewardship, standards, data ownership, and the governance framework. You own the models those policies govern and the platform design that makes them enforceable.
Data Product Management: Partner owns what gets built, for whom, and in what order; the data product backlog and outcomes. You own how it is modeled, what reuse it must honor, and the architecture it must land on.
Data Engineering & Platform: Partner owns pipelines, infrastructure, orchestration, delivery, and the agent runtime. You own the framework, standards, and target model they build against, as well as the design review of what ships.
Data Science & ML: Partner owns models, features, experiments, and evaluation. You own the feature-ready and retrieval-ready substrate, and the environments ML runs in.
Application Teams: Partner owns the systems of record for data created in the organization. You own alignment on data movement patterns and publish/subscribe models, data lineage, and quality frameworks.
You influence without line authority. Engineers, analytics engineers and the global delivery partners supporting the brand and IMG estates adopt your standards because the reference implementations work and the decisions are written down. You raise architectural issues early and in writing, including unpopular ones, and own the trade-off conversation with engineering and business leadership. You explain to non-engineers what a decision costs, what it buys and what it forecloses. You mentor engineers in dimensional modeling, semantic design and master data practice as a discipline distinct from pipeline engineering.
What success looks like
By 6 months. The target-state framework is published and adopted as the standard for the platform unification program: ingest, curation, serving and ML patterns; environments and promotion path; modeling standards; the architecture decision record practice. The first conformed dimensions (fan, event, date, venue, brand) are in production and used by more than one domain. The semantic metric tiers and certification process are defined and the first enterprise-certified metrics are live in the agent estate.
By 12 months. One 360 has a usable production slice: a resolved entity graph with defensible merges, consent lineage, and at least one cross-brand question answerable through governed definitions rather than hand-built SQL. Lineage, catalog coverage and quality checks are automated for every certified data product. Each domain migrating onto the unified platform has a target model defined before its wave begins, and parity criteria that go beyond a green pipeline.
By 18 months. All three 360s have production increments, and the fifth domain onboarded demonstrably cost less than the first. The global delivery partners build to the standard without escalation. Executive-facing numbers from conversational agents trace to certified definitions, and the failure mode for an unanswerable question is a refusal rather than a plausible wrong answer.
What you bring
Must have
- 12+ years in data architecture and engineering, with a substantial portion as a named architect accountable for enterprise-wide design across multiple business units or operating companies, adopted without line authority over the teams adopting it.
- Deep dimensional modeling expertise at enterprise scale: conformed dimensions, grain discipline, slowly changing dimensions, and the judgment to know when a star schema is the wrong answer.
- Hands-on master data architecture across multiple domains: implementation-style selection, entity resolution, survivorship and golden-record design, including what breaks at scale and what it costs to unmerge.
- Production ownership of a semantic layer used by more than one team: metric definition ownership, certification, and the organizational work of collapsing competing definitions of the same number into one.
- Deep, current Snowflake architecture experience: medallion design, role and access taxonomy, tag-driven masking and row-access policies, and the cost consequences of modeling decisions.
- Strong dbt: project and mesh structure, macros, contracts, versions, testing, and the judgment to consolidate overlapping projects.
- Experience architecting through a platform migration or consolidation of real scope, including the parity validation that determined whether it succeeded.
- Practical command of governance from a semantics perspective - glossaries, metric definitions, lineage, data contracts, classification and access models - and how they become enforceable in a platform.
- Exceptional written communication. This role is judged on design documents, standards and decision records that hold up after the author has moved on.
Strongly preferred
- Direct experience making enterprise data consumable by AI systems: semantic models exposed as agent tools, retrieval-ready and context-rich datasets, and a clear view of where generated SQL is appropriate and where it is not.
- Identity graph, Customer 360 or Fan 360 design experience, and working knowledge of consumer privacy regimes (CCPA, GDPR) and consent management as a modeling input.
- Designing the data substrate for ML: feature stores, training and retrieval pipelines, in-platform serving.
- Experience in media, sports, entertainment, live events, ticketing or consumer and fan data environments.
- Familiarity with open semantic interchange, open catalog and metadata standards, and the vendor landscape around them.
- Experience working alongside offshore and partner delivery teams and raising their output to an architecture's standard
Per local requirements and in the interest of transparency, the hourly rate shown below reflects the prevalent current hiring range for this position. Hiring pay rates are based on a number of factors, including location and may vary depending on job-related qualifications, knowledge, skills and experience. The company strives to provide locally competitive rewards packages, which include base rate along with, as applicable, short- and long-term incentives, growth and developmental opportunities, and robust benefits, such as health care, retirement, vacation and other paid time off, and additional offerings.
Hiring Rate Minimum:
$187,500 annually (minimum will not fall below the applicable State/local minimum salary thresholds)Hiring Rate Maximum:
$250,000 annuallyTKO is an Equal Opportunity Employer and complies with all applicable federal, state, and local laws regarding non-discrimination in employment. TKO makes employment decisions based on merit and qualifications, without considering an employee’s or applicant’s race, color, religion, sex, sexual orientation, gender identity or expression, national origin, age, disability, marital status, veteran status, or any other basis prohibited under federal, state or local laws governing non-discrimination in employment in every location in which the Company has facilities. TKO also provides reasonable accommodations for qualified individuals with disabilities in accordance with the Americans with Disabilities Act (ADA) and applicable state or local laws. For information about Privacy and Information Security for TKO employment candidates, please review our Privacy Policy. For information regarding Terms of Use for this and other TKO websites, please review our Terms of Use.

