Parsing Engineer (AI) | Senior

Location: Chantilly, VA
Security Clearance: Active TS or higher [Required]
Job Type: Full-Time
Target Salary Range*: $172,000 - $196,000

*This represents the potential salary range for this position depending on education level, years of experience and/or certifications in addition to other position specific requirements which may impact salary

The Senior Parsing Engineer (AI) builds, optimizes, and maintains text extraction and chunking pipelines that feed unstructured data into large language model (LLM) ecosystems, vector databases, and retrieval-augmented generation (RAG) frameworks. This role preserves document structure and semantic context while transforming complex, multiformat content into clean, Markdown-formatted text. Working with AI researchers, data engineers, and ML platform teams, the engineer improves embedding accuracy, retrieval precision, and context window utilization while minimizing downstream model hallucinations.

• Build and maintain high-fidelity extraction workflows for PDFs, charts, presentations, and web documents.
• Preserve semantic layout and structural metadata, including hierarchies, headers, and bulleted lists.
• Transform raw, unstructured content into clean, Markdown-formatted text for downstream AI processing.

• Implement parent-child document relationships and sliding-window chunking strategies.
• Develop multimodal, vision-based table extraction using tools such as Docling or layout-aware open-source libraries.
• Write specialized code to parse and extract data embedded within tables.

• Enrich text chunks with metadata tags, including source lineage, temporal data, and classification levels.
• Prepare parsed content for LLM ecosystems, vector databases, and RAG frameworks.
• Support high-precision hybrid search through context-preserving extraction and metadata enrichment.

• Benchmark and tune parsing pipelines to improve retrieval accuracy and context window utilization.
• Optimize extraction and chunking workflows to maximize embedding accuracy and minimize downstream model hallucinations.
• Collaborate with AI researchers, data engineers, and ML platform teams to refine pipeline performance.

• Bachelor’s degree required.

• Minimum of 8 years of experience.

• High-fidelity text extraction and document chunking pipeline development.
• Processing of complex, multiformat content while preserving semantic layout and structural metadata.
• Transformation of unstructured data into Markdown-formatted text.
• Parent-child document relationships and sliding-window chunking.
• Multimodal, vision-based table extraction using Docling or layout-aware open-source libraries.
• Specialized coding for table parsing and data extraction.
• Metadata tagging for source lineage, temporal data, and classification levels.
• Parsing pipeline benchmarking and optimization for embeddings, hybrid search, RAG, and LLM context window utilization.