Education - Intelligent Product

Job post
Senior AI / Full-Stack Engineer — Education Intelligence Product (Part-time / Project initially)
I am building an AI-powered mathematics learning intelligence product for students, parents and teachers.
The core intelligence/backend already exists and is being actively developed. I am looking for a strong engineer to help turn the existing system into a robust working product—not to build a chatbot wrapper from scratch.
Initial scope
The first phase is focused on building and polishing:
- Parent/student conversational chatbot
- natural multi-turn conversations
- understands context and follow-up questions
- grounded entirely in student records
- evidence/verification behind answers
- conversation persistence
- Student intelligence
- performance and learning patterns
- strengths/gaps
- topic/chapter/unit analysis
- longitudinal trends
- evidence-backed priorities
- Teacher/class intelligence
- identify patterns across a class/cohort
- students needing attention
- common misconceptions/error patterns
- actionable teaching insights
- Action layer
- evidence-backed “what should we do next?”
- interventions/practice recommendations
- eventually connect recommendations to questions/assessments
- Marks/performance prediction
- evidence-based prediction
- uncertainty/confidence
- no black-box claims
The eventual product will become a professional, scalable web application, but this first engagement is not a full website rebuild.
Existing technology
Current stack includes:
- Python / FastAPI
- React / TypeScript
- PostgreSQL / Supabase
- Vercel / Render
- Git/GitHub
- LLM APIs
- existing AI interpretation, evidence selection, verification and conversation-state architecture
The existing codebase has substantial automated testing and an established architecture. The ability to understand and extend an existing codebase is critical.
Who I am looking for
I want a strong product-minded engineer, not simply a prompt engineer.
You should be comfortable with:
- Python/FastAPI
- React/TypeScript
- PostgreSQL
- APIs and authentication
- LLM/agent applications
- structured outputs and tool calling
- testing and debugging
- Git
- production deployment
- understanding stateful conversational systems
- designing maintainable software rather than accumulating hacks
Education/EdTech experience is a plus.
Very important
I already have a clear product/architecture direction.
I am not looking for someone to throw away the existing architecture and rebuild it using the latest AI framework.
I value:
- clean contracts
- deterministic behaviour where appropriate
- evidence/grounding
- auditability
- security
- maintainability
- minimal architectural changes
- good product UX
You should be comfortable challenging decisions with technical reasoning, but also working within an established architecture when it is sound.
Engagement
Initially:
Part-time / project-based
I would prefer to start with a paid small trial project before committing to a larger engagement.
If the fit is strong, this can grow into a longer-term role covering the complete product and eventual production-scale architecture.
To apply
Please send:
1. 2–3 examples of AI products you have actually built.
2. One example where you extended an existing codebase rather than starting from scratch.
3. Your experience with LLM conversational applications.
4. Your preferred stack.
5. Your hourly/monthly expectation.
6. GitHub/portfolio if available.
7. Briefly explain how you would approach building a grounded student/teacher AI system where the AI must not invent facts from student records.
Please don't send a generic agency proposal or a list of AI technologies. I am looking for an individual engineer who can actually work hands-on in the codebase.