ScholarMind AI / Merit Point Academy
My role
I designed and shipped ScholarMind AI / Merit Point Academy as an independent AI product engineer. I owned the product loop from requirements and architecture through agent-assisted implementation, testing, privacy and attribution guardrails, deployment, and production iteration.
What I built
- A bilingual LLM research assistant.
- A 34-course library with recorded lessons and online learning materials.
- A short-video publishing workflow.
- Robotics and research evidence pages.
- A LeetCode knowledge-graph game with 34 concept nodes and 102 progression units.
- Claude- and self-hosted Qwen/Ollama-backed workflows deployed across multiple subdomains.
Why it matters for forward-deployed engineering
The project demonstrates the same delivery pattern as my robotics work: start with an ambiguous user need, define a workable system, integrate multiple components, ship it to real users, observe failures, and improve the deployed product.