LRNKI / NKG-01STATUS · Active product R&D
Adaptive learning · Active product R&D
lrnki
EVIDENCELive product entry and active product R&D
Visit the live product - The problem
- AI learning tools can produce plausible explanations without exposing how concepts, sources and prerequisite order relate.
- My role
- I lead the product strategy and build the knowledge-graph pipeline, AI evaluation system and full-stack learning product.
- Key decision
- Keep concepts, sources and prerequisite structure inspectable instead of hiding them inside model output.
- Outcome
- A live product that turns curated material into an explorable graph and uses that structure to guide a learner through a difficult topic.
- TypeScript
- Next.js
- Postgres
- Knowledge graphs
- LLM evaluation