
Information Extraction Engineer
Posted 1 day ago

Posted 1 day ago
This is a fully remote position, open to applicants in Virginia.
β’ Extract structured information from lengthy, free-text, narrative, and semi-structured documents.
β’ Ensure a traceable connection from each extracted fact back to the exact source passage from which it originated.
β’ Confirm that every referenced passage genuinely supports the extracted fact, treating any missing or unsupported citation as a defect.
β’ Differentiate between a stated total and a derived sum in financial narratives, while preserving the contributing items and the calculations for verification.
β’ Utilize the provided rule context for model-assisted extraction, keeping cited facts and uncertainties distinct from any validity determination.
β’ Consistently apply and enhance the team's established extraction methods across new document types.
β’ Navigate ambiguous, contradictory, or incomplete source text without forcing a singular interpretation; flag it instead of making assumptions.
β’ Collaborate with entity resolution and retrieval colleagues regarding the requirements of usable extracted facts for downstream processes.
β’ Produce extracted facts at scale, ensuring each claim is traceable to its source passage.
β’ Maintain throughput across various new document types without needing to re-derive the extraction approach for each.
β’ Over 3 years of experience in extracting structured facts or data from unstructured, narrative text in a production environment.
β’ Proven track record of keeping extracted output linked to its source.
β’ Proficient in Python, with practical experience in NLP or information-extraction tools, whether they are rule-based, statistical, or LLM-based.
β’ US Citizenship is Required.
β’ Active US Secret clearance.
β’ Preferred: Experience in extraction within government, investigative, legal, financial-crime, or intelligence-adjacent contexts.
β’ Preferred: LLM-based extraction experience with structured output schemas and validation against source text.
β’ Preferred: Background where extracted facts contributed to a downstream matching, scoring, or retrieval system.
β’ Preferred: Active Top Secret clearance.
β’ A fully remote, results-oriented work environment.
β’ Competitive salary, bonus, and equity package.
β’ 100% employer-covered, comprehensive health insurance, including medical, dental, and vision for you and your family.
β’ Unlimited paid time off, subject to manager's approval.
β’ Flexible work environment allowing you to manage your workday effectively.
β’ 14 weeks of fully-paid parental leave.
CodiLime
Coderio
Get handpicked remote jobs straight to your inbox weekly.