Remote Principal Engineer - AI & LLM Systems

7Seventy Recruiting; United States; On-site; USD 180,000 - 230,000; Full time

Job summary

7Seventy Recruiting seeks a Principal Software Engineer, AI to design, build, and scale production software that combines large language models with deterministic processing. You will own the architecture and help develop an AI-enabled audit workbench that turns raw evidence into accurate, reviewable workpapers through a pipeline-based system.

You will build ingestion/extraction pipelines, recomputation logic, workpaper generation, evaluation infrastructure, and frontend review interfaces.

Qualifications

• Experience building production LLM systems where accuracy is critical.
• Hands-on work with AI extraction pipelines or comparable production LLM apps.
• Strong software engineering abilities across design, implementation, deployment, and operation.
• Experience building evaluation frameworks and detecting regressions before users.
• Ability to balance probabilistic AI with deterministic functions for financial accuracy.

Responsibilities

• Design, develop, and scale an AI-enabled audit workbench processing engagement evidence into structured workpapers.
• Build and maintain document ingestion and information extraction pipelines.
• Develop deterministic recomputation and exception-detection logic to validate extracted data.
• Implement workpaper generation for consistent, traceable outputs for review.
• Design evaluation infrastructure for production LLMs, including datasets and regression testing.
• Create frontend review interfaces for structured human validation and approvals.
• Define boundaries between stochastic AI outputs and deterministic calculations for compliance.

Skills

LLM systems

Software architecture

End-to-end development

Deterministic logic

Data validation

Audit workbench

Regulatory compliance

Frontend interfaces

Production ownership

Code quality

Job description

7Seventy Recruiting seeks a Principal Software Engineer, AI to design, build, and scale production software that combines large language models with deterministic processing. You will own the architecture and help develop an AI-enabled audit workbench that turns raw evidence into accurate, reviewable workpapers through a pipeline-based system.

You will build ingestion/extraction pipelines, recomputation logic, workpaper generation, evaluation infrastructure, and frontend review interfaces.