AI Math Tutor Development

Job ID: 39556914

Budget: $250 – $750 USD

Build Two Versions of AI Math Tutor (Small Language Models, Step-by-Step Teaching)

We are seeking a developer to create two fully working versions of a server-hosted AI math tutor using different small language models (SLMs) — each under 7B parameters.

Both versions will teach math in a step-by-step, Socratic style: asking questions, encouraging reasoning, checking for understanding before explaining or moving on.


Scope of Work:
Deploy two different open-source SLMs (GGUF or Hugging Face-compatible) on our server or GPU cloud
Build a simple chat-based tutor interface (can be CLI or web-based)
Design custom prompts to make the tutor:
-Guide students through math problems step-by-step
-Ask questions instead of giving immediate answers
-Explain reasoning in a warm, encouraging tone
-Pause and check for understanding
Allow switching between the two models for comparison
Provide installation scripts and clear documentation

Requirements:
Experience deploying small LLMs (e.g., via Ollama, vLLM, llama.cpp, or similar)
Familiarity with prompt engineering or tutoring-style prompts
Python-based deployment (Flask, FastAPI, or similar)
Dockerized setup preferred

Deliverables:
Two running AI tutor agents using different SLMs
Hosted chat or test interface
Setup scripts (Docker preferred)
System prompt customization guide

Apply If You:
Have experience hosting small models locally or on cloud
Can deliver working demos within 10–14 days
Can recommend the best SLMs for reasoning and math guidance