Software Developer (Backend – Integration)

Job Family:

Data Science & Analysis

Travel Required:

Up to 10%

Clearance Required:

Active Top Secret (TS)

Guidehouse is seeking a Software Developer to join our Technology / AI and Data team, supporting mission-critical initiatives for Defense and Security clients. In this role, you will lead the design and implementation of secure, scalable ingestion and data processing workflows that power advanced AI-driven platforms. You will architect solutions for transforming complex, high-volume data into structured outputs optimized for downstream AI/ML pipelines, while ensuring compliance with stringent federal security and regulatory standards. Collaborating with engineers, architects, and mission stakeholders, you will deliver innovative backend capabilities that enable accurate, efficient, and reliable decision-making in support of national security objectives.

What You Will Do:

• Serves as the lead backend integration engineer responsible for architecting and implementing ingestion, preprocessing, normalization, and transformation workflows for the FBI adjudication AI platform.
• Designs ingestion frameworks supporting SF-86 forms, investigative attachments, summaries, financial/criminal records, and continuous vetting alerts using both traditional OCR and VLM/LLM-based document understanding.
• Ensures ingestion workflows comply with FedRAMP High, RMF, CJIS, and FBI ATO requirements, including logging, auditability, encryption, and secure processing of PII and sensitive investigative information.
• Collaborates with AI/ML engineers, backend API developers, cloud engineers, and security engineers to ensure ingestion outputs are optimized for RAG workflows, SEAD-4 scoring, anomaly detection, and adjudicator review.

• Data Ingestion, Parsing & ETL Architecture

• Design ingestion pipelines supporting LLMs and VLMs for OCR, document understanding, multimodal extraction, and parsing of complex investigative materials including forms, tables, handwritten elements, and embedded imagery.
• Build scalable ingestion and ETL workflows capable of processing hundreds of pages per case using OCR engines (Textract, Tesseract) and VLM-based parsing models such as LayoutLM, Qwen-VL, Donut, or LLaVA.
• Implement normalization and transformation workflows including deduplication, schema harmonization, field mapping, classification labeling, chunking, segmentation, and tokenization optimized for downstream LLM/RAG operations.
• Develop fault-tolerant ingestion systems with checkpointing, idempotency, retry frameworks, ingestion-state tracking, and structured error reporting.

• Backend Integration & System Connectivity

• Build secure, compliant integrations with FBI systems, case repositories, identity/HR systems, and continuous vetting alert sources using APIs, ETL endpoints, SFTP, and message queues.
• Develop backend microservices that assemble case packages, correlate evidence across disparate sources, and produce structured adjudication-ready datasets.
• Integrate ingestion outputs with vector databases, embedding pipelines, and LLM inference services, ensuring data is structured, enriched, and optimized for reasoning workflows.
• Ensure all integrations enforce strict authentication, authorization, validation, and data-handling policies.
• RAG / LLM Data Preparation
• Create ingestion workflows that prepare documents and extracted content for embeddings, retrieval indexing, semantic search, and long-context reasoning.
• Implement chunking, segmentation, labeling, and evidence-tagging strategies designed to maximize retrieval precision and reduce hallucination risk in LLM inference.
• Develop heuristics for filtering, prioritizing, and contextualizing extracted information to enable fact-grounded SEAD-4 scoring and memo generation.
• Support preparation of vector representations, metadata fields, and retrieval keys for large-scale evidence collections.

• Security, Compliance & Logging

• Implement secure ingestion pipelines aligned with FedRAMP High, RMF, CJIS, and FBI security requirements including encryption, access control, PII-handling rules, and secure logging.
• Apply advanced PII-safe processing techniques including automated redaction, VLM-aided sensitive field detection, classification tagging, and compliance-driven filtering.
• Ensure ingestion systems generate detailed logs, lineage metadata, provenance trails, and audit events supporting adjudication oversight and accreditation documentation.
• Collaborate with Security Engineers to ensure ingestion controls map to SSP requirements and POA&M items are remediated promptly.

• Performance Optimization & Reliability

• Optimize ingestion pi