Senior Full-Stack AI Developer for Education Platform

We need an AI technician who can help with building ScolaNova, an AI-powered education platform for Académie de Reims, that helps schools manage curriculum, support teachers, improve student learning, and automate administrative work. The platform is already functional and includes separate areas for administrators, teachers, students, parents, and applicants. We’re now looking for an experienced Full-Stack AI Developer to join our project team for the next development stage. What is already built The current platform includes:

• School administration and multi-role portals
• Curriculum, lessons, assessments, and report cards
• Admissions, billing, and school websites
• AI-generated lesson plans, assignments, question banks, and assessment drafts
• Curriculum extraction from uploaded documents
• PDF text extraction and OCR
• A School Knowledge Assistant with document retrieval and citations
• Server-side AI integration through Convex and OpenRouter What we need help with Your first task will be to review the current codebase and AI implementation. We want to understand what is working well, what needs improvement, and how to complete the next stage without replacing reliable features. The main focus will be:
• Improving document search and RAG quality
• Strengthening embeddings, retrieval, and source citations
• Making sure users only access permitted school documents
• Improving AI reliability, testing, monitoring, and error handling
• Connecting curriculum requirements with lessons and assessments
• Preparing the foundation for a Teacher Copilot
• Supporting future administrative automation and student-learning features We want someone who can work across the frontend, backend, data, and AI layers. This is an applied product-development role, not an AI research position. Current technology
• Next.js
• React
• TypeScript
• Convex
• OpenRouter
• LLM and RAG workflows
• PDF extraction and OCR
• Role-based, multi-tenant architecture Experience with Convex is helpful but not required if you are comfortable learning an existing backend architecture. Who would be a good fit? You may be a strong fit if you have experience with:
• Building production applications with Next.js and TypeScript
• Integrating LLMs into real products
• RAG, embeddings, vector search, and grounded answers
• Document processing, PDF extraction, or OCR
• Backend APIs, authentication, and permissions
• Reviewing and improving an existing codebase
• Testing AI behavior and solving retrieval-quality problems Education technology experience would be helpful, but it is not required. We are more interested in practical project experience than a long list of technologies. You should be able to explain what you personally built and how it worked in production. Initial collaboration This is initially a short-term, project-based role. The first stage will include: 1. Reviewing the existing platform 2. Identifying technical and AI-related gaps 3. Creating a practical implementation plan 4. Improving one complete AI workflow 5. Testing and documenting the completed work 6. Recommending the next development priorities If the collaboration goes well, there may be opportunities to support later product stages. Compensation: €50–€70 per hour, depending on experience and availability. The initial engagement is expected to last 6–10 weeks at approximately 20–30 hours per week. We are also open to agreeing on milestone-based pricing after the initial technical review. How to apply Please send us:
• A short introduction
• Examples of relevant AI or full-stack projects

… lees de volledige omschrijving bij Freelancer.nl.