AI Tutor for Personalized Lessons
Budget: $2 – $8 USD
I am building an AI-powered tutor whose main job is to deliver fully personalized lessons rather than simply answering isolated questions. The agent should be able to:
• analyse a learner’s current knowledge from a short diagnostic prompt,
• generate a step-by-step lesson sequence that adapts in real time to every response, and
• store session history so future lessons pick up exactly where the student left off.
I expect the core to run on a large language model such as GPT-4 (other LLMs are fine if they meet the quality bar) combined with a lightweight database or vector store to keep user profiles and lesson context. A simple web interface is enough for now; clean API endpoints are essential so I can later add mobile clients and progress-tracking dashboards.
Deliverables I need from you
• A working prototype accessible through the browser
• Source code with clear inline comments and a short setup guide
• Documentation explaining how personalization is achieved and where to extend the curriculum
Completion is defined by my ability to log in as a test learner, select a topic from the sample set I will provide, and receive a coherent, adaptive lesson that changes when I intentionally give wrong or right answers.
When you reply, please outline the toolchain you prefer (e.g., Python, LangChain, Pinecone, Next.js) and share one example of a similar adaptive-learning feature you have shipped.
• analyse a learner’s current knowledge from a short diagnostic prompt,
• generate a step-by-step lesson sequence that adapts in real time to every response, and
• store session history so future lessons pick up exactly where the student left off.
I expect the core to run on a large language model such as GPT-4 (other LLMs are fine if they meet the quality bar) combined with a lightweight database or vector store to keep user profiles and lesson context. A simple web interface is enough for now; clean API endpoints are essential so I can later add mobile clients and progress-tracking dashboards.
Deliverables I need from you
• A working prototype accessible through the browser
• Source code with clear inline comments and a short setup guide
• Documentation explaining how personalization is achieved and where to extend the curriculum
Completion is defined by my ability to log in as a test learner, select a topic from the sample set I will provide, and receive a coherent, adaptive lesson that changes when I intentionally give wrong or right answers.
When you reply, please outline the toolchain you prefer (e.g., Python, LangChain, Pinecone, Next.js) and share one example of a similar adaptive-learning feature you have shipped.