5+ years of professional software development experience Strong experience with Node.js and TypeScript Experience building production systems using serverless functions Strong backend-focused full- instincts, with comfort supporting web application workflows end-to-end Experience implementing authentication, OAuth, session handling, and role-based access control (RBAC) Strong understanding of relational databases, particularly PostgreSQL Ability to design reliable backend logic involving user decisions, workflow state transitions, and auditability Disciplined Git and code-review workflow Strong capabilities in prompt-driven development using AI tools (like Cursor or Copilot) to scaffold architecture and plan work, while maintaining strict code quality discipline Understanding of state transitions, workflow orchestration, and deterministic vs. AI-driven decisions Comfortable with serverless quirks like stateless functions, cold starts, idempotency, and retry behaviour Nice-to-Have: Familiarity with Python, Linux, and web server administration (AWS EC2) Experience with vector search approaches like pgvector, vector databases, FAISS, and BM25 search matching Experience integrating LLM-powered features and retrieval-enabled (RAG) applications Frontend capability (React / Next.js) for building clean, usable admin and workflow interfaces, as well as familiarity with Gradio Familiarity with Supabase or Firebase Auth for user management Knowledge of observability and debugging tools for AI systems (e.g., Langfuse) to trace decisions and failure modes Familiarity with Server-Sent Events (SSE) or websockets for real-time UI updates during long-running processes Experience with MCP-style tool exposure, JSON contracts, and chaining steps (plan → execute → validate) Experience using queues with serverless architectures (e.g., SQS, Pub/Sub) for async job handling Responsibilities: Build and maintain serverless Node.js functions that power the core application workflows Develop AI-assisted document processing pipelines involving ingestion, normalisation, classification, extraction, and retrieval-enabled features Implement structured workflows handling document review, compliance checks, exception handling, and audit trails Support systems where authorised users can override or make case-by-case decisions that affect downstream workflows Track user actions, approvals, and decision history in a reliable and auditable format Integrate authentication and role-based permissions using providers like Supabase or Firebase Build or support internal admin interfaces for operational oversight Collaborate heavily on data modelling, schema design, backend architecture, and secure product logic Define practical engineering patterns for prompt-based development and code maintainability Design "human-in-the-loop" systems, deciding precisely when AI should assist, when strict rules sho