Remote Information Extraction Engineer - Traceable Facts

Openkrill; United States; À distance; USD 130 000 - 160 000; Plein temps

Résumé du poste

DEFCON AI is seeking a remote data extraction specialist to pull structured facts from long, free-text documents within a secure government cloud environment. You will ensure every fact has a precise source citation and maintain traceability for reviewer verification.

You will work with an analytics and AI engineering team, supporting high-volume extraction, validation, and defensible output across diverse document types, while flagging ambiguous or unsupported passages for review.

Qualifications

• 3+ years extracting structured facts from unstructured narrative text in production settings.
• Experience keeping extracted output traceable to its source.
• Strong Python, with hands-on NLP or information-extraction tooling, rule-based or statistical or LLM-based.

Responsabilités

• Extract structured facts from long, free-text narratives and ensure citations back to source passages.
• Maintain a traceable pointer from every extracted fact to the exact source passage.
• Validate that each cited passage supports the extracted fact and flag unsupported citations as defects.
• Distinguish a stated total from a derived sum in financial narratives and retain contributing items and calculations.
• Use and extend the team's extraction approach across new document types with consistency.
• Handle ambiguous, contradictory, or incomplete source text without forcing a single reading; flag issues.
• Collaborate with entity resolution and retrieval teammates to ensure downstream compatibility.

Connaissances

Python

NLP

Data extraction

Traceability

Attention to detail

Outils

LLM-based extraction

Rule-based NLP

Entity resolution systems

Description du poste

DEFCON AI is seeking a remote data extraction specialist to pull structured facts from long, free-text documents within a secure government cloud environment. You will ensure every fact has a precise source citation and maintain traceability for reviewer verification.

You will work with an analytics and AI engineering team, supporting high-volume extraction, validation, and defensible output across diverse document types, while flagging ambiguous or unsupported passages for review.