AI EdTech Platform Developer (Python, LLM/LMM)

1. Title 1.1 AI Platform Developer (Hands-On Build, Python, LLM/LMM Integration – EdTech) 2. Description 2.1 We are seeking a highly capable AI Platform Developer to build an end-to-end AI-powered education platform from scratch. This is a hands-on development role focused on rapid execution, system building, and continuous iteration. We are not looking for a manager or purely architectural role. 2.2 The platform will support Reception to Year 12 and will include AI tutors (Maths, Physics, Sciences, English), AI-based curriculum development, automated resource generation, AI marking systems, and a large-scale question bank platform. The system will initially be used internally and will later evolve into a scalable online learning platform for student self-paced learning. 2.3 This role requires practical experience integrating both LLM (Large Language Models) and LMM (Large Multimodal Models) into real-world applications, including text generation and multimodal outputs such as diagrams and visual content. 3. Scope of Work 3.1 Design and develop backend systems using Python (FastAPI or similar) 3.2 Build AI-powered modules including AI tutor, curriculum generator, resource generator, marking system, and question bank system 3.3 Integrate multiple AI APIs (OpenAI, Claude, Gemini or similar) into a unified codebase 3.4 Develop systems using both LLM (text generation and reasoning) and LMM (multimodal outputs such as diagrams and visuals) 3.5 Design structured prompt engineering workflows and JSON-based outputs 3.6 Build scalable APIs for internal tools and future platform integration 3.7 Design and implement database schemas for questions, curriculum, student progress, and analytics 3.8 Develop batch systems for large-scale content generation (10,000+ questions and resources) 3.9 Ensure clean, modular, and maintainable code 4. AI Development Lifecycle Responsibilities 4.1 Define the problem and system objectives for each AI module 4.2 Design and implement data structuring and preparation workflows 4.3 Develop AI-driven systems using LLM and LMM integrations 4.4 Evaluate and test outputs for accuracy, consistency, and curriculum alignment 4.5 Deploy and integrate modules into a unified backend system 4.6 Monitor system performance and output quality 4.7 Continuously improve prompts, workflows, and system performance 5. Tech Stack 5.1 Python (FastAPI preferred) 5.2 PostgreSQL or similar relational database 5.3 AI APIs (OpenAI, Claude, Gemini or similar) 5.4 Experience working with both LLM and LMM systems 5.5 Docker (preferred) 5.6 Cloud infrastructure (AWS, GCP, or Cloudflare) 5.7 Git and version control 6. Ideal Candidate 6.1 Proven experience building AI applications from scratch 6.2 Strong hands-on backend development skills in Python 6.3 Experience integrating and orchestrating multiple AI models 6.4 Strong understanding of LLM and LMM capabilities 6.5 Experience with structured outputs and prompt engineering 6.6 Ability to build quickly and iterate systems efficiently 6.7 Experience in EdTech or learning systems is highly desirable 7. Project Context 7.1 This is a core technology build for an established education business with a clear roadmap to scale into a global AI-powered learning platform 7.2 The focus is on building high-quality internal systems first, followed by productisation into a student-facing platform 8. Engagement 8.1 Long-term engagement with ongoing development work 8.2 Opportunity to grow into a lead technical role over time 8.3 Direct collaboration with founder on system design and product direction 9. Application Requirements 9.1 Examples of AI systems or applications you have built 9.2 GitHub or portfolio showcasing relevant work 9.3 Brief explanation of how you would approach building an AI-powered education platform with LLM and LMM integration 9.4 Your experience working with structured outputs and multi-model AI systems 10. Screening Question 10.1 How would you design a system that generates high-quality, curriculum-aligned maths questions at scale, ensures accuracy, avoids duplication, and supports diagram or visual generation where required?

Projektin tunnus (ID): 40321877