Senior Data Platform Engineer

DEFCON AI is an insights company that leverages artificial intelligence, mathematical optimization, data analytics, and software engineering to optimize complex systems and improve resilience. The Senior Data Platform Engineer will own and build a durable storage layer that preserves source disagreement, provenance, uncertainty, and historical decisions while supporting reproducible, point-in-time analysis.

Responsibilities:

• Design and build the storage layer that preserves source disagreement and history rather than resolving them away at write time
• Build the first release’s result and evidence store: saved per-person results linked to the originating record, prior results and human feedback retained separately, and configuration versions on every result
• Implement provenance and lineage across every node and edge against the platform’s evidence-record contract
• Store the versioned rule library and configuration, and record on every result the exact rule versions and ordered context supplied to any model-assisted step
• Ensure every write is traceable to its origin and every decision is reproducible
• Carry match confidence and other uncertainty forward through the platform rather than collapsing it into a single value
• Support point-in-time questions: what did we believe about this entity on this date, and on what evidence
• Define, with entity resolution and ML teammates, what the storage layer needs to receive and what it must serve downstream

Requirements:

• 6+ years in data engineering or data platform engineering, including production experience with versioned, temporal, or historized data models
• Experience designing storage where history, provenance, and source disagreement are first-class rather than resolved away
• Comfort building on probabilistic upstream output rather than a clean shared key
• Strong Python and SQL, with production experience on PostgreSQL or comparable
• US Citizenship Required
• Active US Secret clearance
• Experience building the storage layer downstream of an entity-resolution or record-linkage system
• Bi-temporal, event-sourced, graph, or temporal data modeling in production
• Federal, government, or regulated-industry experience where a historical decision had to be reproducible on demand
• Active Top Secret clearance