Just Now Possible Inside eSpark’s AI Teacher Assistant: RAG, Evals, and Real Classroom Needs
15 snips
Sep 25, 2025 Thom van der Doef, Principal Product Designer at eSpark, and Ray Lyons, VP of Product & Engineering, dive into the fascinating development of the AI Teacher Assistant. They discuss the evolution from a chatbot interface to a structured workflow tailored for educators. Key insights include how retrieval augmented generation (RAG) and refined metadata are optimizing the assistant's performance, plus lessons learned in semantic versus keyword search. They also explore future enhancements using student data for personalized recommendations, aiming to bridge teacher mandates with real classroom needs.
AI Snips
Chapters
Transcript
Episode notes
LLMs Make Curriculum Integration Scalable
- eSpark used admin and teacher interviews to find a cross-stakeholder opportunity tying supplemental content to district core curricula.
- LLMs made scalable matching feasible where manual pairing of many curricula and activities would not.
Chatbot Prototype Flopped With Teachers
- The team initially prototyped a conversational chatbot with an open text box and watched teachers freeze at the blank prompt.
- Multiple teacher interviews showed open text chat was too slow and unfamiliar for teachers' time-pressed workflows.
Prefer Guided Choices Over Free Chat
- Present constrained choices (select lesson → recommendations) instead of open text to reduce friction and guessing.
- Add three contextual follow-up suggestions so teachers can take one-click next steps and you can learn which paths gain traction.