AI in Education

The Role of Custom LLMs as Personalized AI Tutors

Personalised AI tutors are now a credible product, but generic LLMs hit a ceiling fast. Here is what changes when you build the model around the curriculum.

2 min read
Custom LLMs acting as personalised AI tutors

Generic LLMs make impressive demos. They make less impressive AI tutors — too eager to answer when they should ask, too generic when learners need scaffolding tied to their curriculum.

Here is how we build personalised AI tutors that meaningfully change learning outcomes.

#1Why generic models hit a ceiling in education

Education is contextually dense. A geometry tutor needs to know the sequencing of theorems in this particular curriculum, the misconceptions the student has shown this week, and the tone the school has approved. A generic model does none of those by default.

#2Curriculum as data, not just content

We treat the curriculum as a structured graph: skills, prerequisites, common misconceptions, exemplar solutions, and assessment items. That graph becomes the retrieval substrate for the model, ensuring every answer is grounded in the school's actual material.

#3RAG vs. fine-tuning vs. both

Retrieval-augmented generation gives the model fresh, curriculum-specific context at inference time. Fine-tuning gives the model the right voice, scaffolding behaviour, and refusal patterns. Most production tutors need both layered carefully — RAG for facts, fine-tuning for pedagogy.

#4Evaluating an AI tutor honestly

Evaluation is the hardest part of EdTech AI. Accuracy metrics miss the point — a tutor can be 'correct' and pedagogically harmful (by giving away answers). We pair automated evaluation with classroom A/B trials and qualitative review by teachers to assess whether the tutor actually moves learners forward.

The takeaway

A personalised AI tutor is a product, not a prompt. It compounds curriculum data, pedagogical fine-tuning, and trustworthy evaluation into something that earns a teacher's recommendation.

Frequently asked questions

Do we need to train our own model?
Rarely. Most production tutors fine-tune an existing foundation model and pair it with curriculum-aware retrieval. Training from scratch is expensive and seldom necessary.
How do we keep an AI tutor safe for younger learners?
Layered safety: input filters, retrieval scope limits, response filters, and human-in-the-loop review. Treat safety as a product surface, not a single setting.
Keep reading

Similar articles