Staff Full-Stack AI Engineer

The Staff Full-Stack AI Engineer (Level IV)is a senior technical leader within AssistRx Engineering. This is a software engineering role with an AI focus, not a narrowly specialized machine learning role. The engineer designs, builds, and deploys end-to-end production solutions across the specialty medication access ecosystem, including patient onboarding, benefits verification, prior authorization, document intake, and patient and provider engagement.

Depending on the business problem, the solution may include a modern web application, REST APIs and backend services, integrations, data workflows, cloud infrastructure, generative AI, predictive models, intelligent document processing, agentic automation, or a combination of these. The Staff Engineer is expected to select the simplest effective approach and to use conventional software or deterministic logic when AI is not the right answer.

The role partners closely with Software Engineering, Product, Architecture, Data, Security, Compliance, and Client Services. It owns complex cross-system solutions, establishes reusable engineering patterns, mentors other engineers, and is accountable for solution quality, reliability, security, maintainability, model performance, cost efficiency, and the safe and compliant use of protected health information.

Responsibilities

Full-Stack Solution Design And Delivery

• Lead the design, development, testing, deployment, observability, and production support of complete applications spanning user interfaces, APIs, backend services, integrations, data access, cloud services, and AI-enabled components.
• Build both GenAI functionality and conventional application functionality based on the needs of the initiative; do not limit ownership to the AI service layer.
• Translate ambiguous business and healthcare workflow requirements into clear technical strategies, designs, incremental releases, and measurable production outcomes.
• Design secure, maintainable REST APIs, event-driven services, system integrations, and responsive web experiences that integrate naturally with existing AssistRx platforms.
• Resolve cross-team dependencies and take proven pilots through production hardening and scale across products and clients.
• Set standards for CI/CD, source control, containerization, and infrastructure-as-code practices for AI workloads
• Serve as the escalation point for complex production issues and client-reported defects in AI-enabled features, leading root cause analysis and continuous improvement of deployed AI services

Architecture And Technology Selection

• Evaluate architectural alternatives and select languages, frameworks, data stores, integration patterns, and cloud services based on business fit, scalability, security, maintainability, cost, and delivery speed.
• Translate ambiguous business and clinical workflow requirements into technical strategies, designs, and production releases, taking proven pilots to scale across products and clients
• Work productively across multiple programming languages and architectural styles rather than defaulting to a single preferred stack.
• Author and review design documentation covering system boundaries, interfaces, data flows, security controls, failure modes, observability, evaluation, and fallback behavior.
• Identify technical debt and platform gaps, then lead pragmatic investments that improve delivery across multiple teams.
• Create reusable components, reference implementations, and engineering standards that accelerate both AI-enabled and conventional application development.

Applied AI Engineering

• Design and implement production LLM applications using retrieval-augmented generation, structured outputs, prompt design, function and tool calling, and agent-based orchestration.
• Develop interactive AI products such as chat, conversational interfaces, voice experiences, assistants, and agentic workflows that are integrated into complete user journeys.
• Design clear system boundaries, memory and state management, routing, human review checkpoints, fallback paths, and deterministic controls for high-impact workflows.
• Select, tune, and evaluate models against defined targets for accuracy, groundedness, hallucination rate, task completion, latency, and cost
• Design agentic workflows with task decomposition, tool and function calling, memory and state management, routing, clear boundaries, human review checkpoints, and fallback paths for high-impact decisions …