Applied AI Software Engineer

Position summary

AI solution design and development

• Translate requirements into secure, maintainable AI designs and technical specifications.
• Develop AI applications, services, agents, and APIs with Python, TypeScript/Node.js, and approved frameworks, models, and cloud services.
• Build RAG solutions, including ingestion, chunking, embeddings, vector search, grounding, prompt design, and orchestration.
• Design multi-stage LLM pipelines and hybrid retrieval with section-aware chunking, reranking, source attribution, resumability, and human approval points.
• Implement schema-validated model outputs and tool calling, and document ingestion, generation, redlining, and change summaries across Word, PDF, and web sources.
• Create reusable components and documentation; participate in architecture, design, and code reviews.

Engineering and delivery

• Build and maintain data pipelines, model integrations, REST APIs, and application components.
• Apply version control, automated testing, peer review, CI/CD, containerization, and infrastructure automation.
• Integrate AI services with enterprise applications, workflows, identity controls, data platforms, and approved cloud environments.
• Develop unit, integration, regression, performance, and safety tests; troubleshoot defects and production issues.
• Build full-stack components with TypeScript, Node.js, React and Next.js or equivalent, and Python, using relational databases with row-level security, versioned migrations, and immutable audit trails.
• Use agentic development practices and coding agents with specification-driven workflows, validation gates, peer review, tests, and source-control traceability.

Responsible AI, security, and quality

• Embed privacy, security, accessibility, human oversight, transparency, and responsible AI requirements throughout development.
• Design evaluation datasets and automated checks for accuracy, relevance, groundedness, hallucination, bias, toxicity, robustness, and task completion.
• Protect health and personal information through approved data handling, access controls, logging, encryption, and secure development practices.
• Document model choices, data lineage, limitations, risks, controls, testing evidence, and operational procedures for governance and auditability.
• Work with privacy, security, legal, clinical safety, and risk stakeholders to resolve issues before release.

Requirements

• Develop applications and services using Python and TypeScript/Node.js.
• Apply software engineering practices across testing, version control, peer review, CI/CD, containerization, and infrastructure automation.
• Integrate privacy, security, accessibility, human oversight, transparency, and responsible AI considerations into development.
• Use human review and traceable testing for AI coding agent-generated work.