Una candidatura completa en un minuto
Currículum y carta de presentación personalizados, listos para enviar.
Neutron Pte. Ltd. is seeking a Full-Stack Engineer to design and build an AI-driven casework platform for the public sector. You will develop Node.js services and TypeScript-based back- end APIs, craft React interfaces for officers, and implement agentic AI workflows with strict audit trails and source traceability.
You will work in a small, agile squad to deliver government-grade security, SSO integration, OCR/document ingestion, and governance with WCAG accessibility on AWS.
As a Full-Stack Engineer, you will design and build the platform end to end, from the Node.js services, agentic AI workflows and system integrations that power it to the React interface officers use every day to review AI-extracted facts, verify them against source documents, and approve reports and recommendations.
Design and build APIs and backend services in Node.js and TypeScript for case intake, document processing, report generation and officer review workflows
Build responsive web interfaces in React and TypeScript for case dashboards, case detail views, document review, task lists, and search, filtering and reporting across large volumes of cases
Build LLM-powered agentic workflows using tool calling and structured outputs to extract key facts, cross-check them, and draft structured reports and recommendations
Build document ingestion and extraction pipelines for digital and handwritten documents, including OCR and handwriting recognition
Implement retrieval over knowledge sources (RAG) and a configurable rules layer that applies business and legal rules to extracted facts
Design review-and-approve workflows where officers check AI-extracted facts, cross-check results, and edit draft reports and recommendations before approving them
Make every AI output traceable to its source, with document viewers that highlight the original digital or handwritten documents and audit trails that record officer review and approval decisions
Set up AI quality evaluation, including test datasets, accuracy metrics and guardrails against hallucination
Integrate with existing enterprise systems via APIs and with government single sign-on (e.g. TechPass / Singpass)
Apply government-grade security and data protection in line with IM8, meet accessibility (WCAG) requirements, and deploy on AWS (GCC) with CI/CD pipelines
Design reusable, configurable services and UI components that other teams in the agency can adapt for their own workflows, in line with government design standards (SGDS)
Work in a small agile squad with designers, product owners and officers, using AI-assisted development to deliver quickly
4+ years of full-stack development experience with Node.js, ReactJS and TypeScript (must have)
Strong experience building RESTful APIs and working with relational databases (PostgreSQL preferred)
Strong hands-on experience with React hooks, component design, HTML, CSS and modern state management in strongly typed TypeScript codebases
Experience building data-heavy dashboards, tables and forms, with a good eye for UX in professional or internal tools
Hands-on experience building LLM applications (OpenAI, Azure OpenAI, AWS Bedrock or similar), including tool calling, structured outputs and prompt design
Experience with retrieval augmented generation (RAG), embeddings and vector search
Experience with AWS, Docker and CI/CD
Good understanding of secure coding practices, data protection and authentication, including single sign-on
Familiarity with automated testing across the stack (e.g. Jest, React Testing Library, Playwright or Cypress)
Comfortable working in an agile squad and using AI-assisted development tools (e.g. GitHub Copilot, Cursor, Claude Code)
Experience on Singapore government projects (GCC, SGTS, IM8, SGDS, TechPass/Singpass integration)
Experience with agentic AI frameworks (e.g. LangGraph, LlamaIndex, Vercel AI SDK)
Experience with document processing, OCR and handwriting recognition (e.g. AWS Textract), and document viewers and annotation (e.g. PDF.js)
Experience building interfaces for AI features such as streaming responses, confidence indicators or human-in-the-loop review
Experience with AI evaluation and observability tools (e.g. Langfuse, Ragas, promptfoo) and guardrails against hallucination
Exposure to rules engines, case management or workflow systems