AI Full Stack Developer

• Internal AI tool delivery — Design, build, and maintain a suite of internal tools on Retool — from initial prototype to production deployment. Own the full stack: Retool application, Python backend API, and LLM integration layer. First-year focus is RCM, billing, and clinical operations workflows.
• LLM prompt engineering & optimization — Author, test, version, and maintain prompts for BHC-specific use cases including denial review, authorization summarization, clinical documentation assistance, and RCM workflow triage. Own the prompt library and its evaluation framework.
• Data layer integration — Connect internal tools to BHC’s Snowflake and dbt data layer, source system APIs (IMS and related), and operational data feeds. Know where the data lives, how to query it efficiently, and how to surface it safely inside a tool.
• Application-layer PHI controls — Implement and maintain HIPAA-compliant data handling within the application layer: PHI masking, role-based access, audit logging, and data minimization patterns. Own application-level security in coordination with the Director and legal/compliance.
• User experience for operational staff — Design interfaces for RCM coordinators, billing staff, and clinical ops users working under time pressure — not for engineers or analysts exploring dashboards. Understand what ‘usable’ means to someone processing 40 authorizations a day.
• Iteration & adoption — Instrument tools for usage tracking, gather structured feedback from operational teams, and ship improvements regularly. Measure adoption and outcomes, not just deployment.
• Documentation & tooling standards — Maintain clear documentation of APIs, prompt patterns, integration architecture, and reusable Retool components. Build the pattern library that makes the second tool faster to ship than the first.

    Qualifications: 

• Strong Python development experience — REST API design and development (FastAPI preferred), async patterns, data validation, and production error handling.

• Hands-on Retool experience at a production level — multi-step workflows, role-based access, API resource configuration, custom components, and Retool-native automation. Equivalent deep experience with a comparable low-code internal tooling platform considered.
• Direct LLM API integration experience — prompt engineering, tool/function calling, RAG implementation, context window management, and response evaluation. Experience with Claude Enterprise, AWS Bedrock, OpenAI, or similar platforms.
• Experience building HIPAA-compliant applications — PHI handling, audit logging, role-based access controls, and data minimization at the application layer. Understanding of what HIPAA requires at runtime, not just policy.
• Practical SQL proficiency — able to query Snowflake or equivalent columnar stores; working familiarity with dbt models as data sources is a meaningful plus.
• Ability to translate operational workflows into application requirements with minimal guidance — comfortable interviewing an RCM coordinator, identifying the actual pain point, and building against it rather than against a spec.
• Healthcare, RCM, or billing domain experience is preferred. Exceptionally strong technical candidates without healthcare background will be considered if they demonstrate fast domain ramp.