Senior Full-Stack Engineer (LLM Decision Support) (Remote)

Job DescriptionWe are building a technology platform focused on LLM-powered decision support. The product combines structured inputs, deterministic business rules, and LLMs used strictly as a support layer (not autonomous agents, not model training). The MVP is intended as a solid foundation, not a disposable prototype.The RoleWe are looking for one senior, autonomous full-stack developer to lead the technical implementation of the MVP, working directly with the founder. This is a lean, execution-focused engagement - not an agency setup, research project, or large team environment.Outcome (2–3 months): deliver a production-ready MVP that runs end-to-end decision flows reliably, is deployed across dev/stage/prod, and is stable enough for a controlled, limited real-user beta.Engagement model: Contract role, anticipated ~2–3 months. Preferred: ~20 hours/week initially. Async-first with structured written updates and 2–3 async check-ins per week; live calls as needed.Core MVP Technical Direction (Defined & Stable)Backend: Node.jsFrontend: ReactDecision & LLM orchestration layer: Python (separate service)Database: Relational (MySQL / PostgreSQL)Architecture: API-first (REST)Within this framework, the selected developer is expected to own implementation details and tooling choices, proposing pragmatic improvements without redesigning the core architecture or expanding scope.Separation of concerns:Node.js owns core business rules and API structurePython service owns LLM calls, prompt structuring/orchestration, and output validationKey Responsibilities:Implement structured input flows and deterministic decision logicBuild and maintain a Python service for LLM orchestration (support layer only)Ensure LLM features support: structured outputs, deterministic prompting, fallback logic, and auditability of inputs/outputsDesign and build clean, maintainable REST APIsIntegrate external APIs where required (expected to be a limited set)Own environment setup (dev/stage/prod), CI/CD, and basic logging/monitoring (keep it pragmatic and lightweight)Deliver production-ready, well-documented code aligned with MVP constraintsMake sound technical decisions aligned with defined scope and timelineRequirements:Strong experience as a senior full-stack developer shipping production applicationsSolid Node.js and React backgroundExperience building and deploying a Python API/service (e.g., FastAPI/Flask)Practical experience integrating LLM APIs in real products (not demos), including reliability patterns:structured outputs + validationdeterministic promptingfallback/retry behavioraudit trails / traceabilityHigh autonomy and product mindset; comfortable owning delivery in a lean environmentScope NotesMust-have:Structured input flowsDeterministic rule layerPython-based LLM orchestration serviceExplainable outputs (clear reasoning path)API-first architectureProduction-ready implementation with dev/stage/prod, CI/CD, and basic observabilityOut of scope:Model training or fine-tuningAutonomous agentsAdvanced admin toolingHeavy analyticsPremature scaling optimization