In developmentIn development, building for production. No public repository yet.
AssesGen
A multi-agent system for automated IELTS question generation and evaluation. Generation, scoring, and feedback run on locally hosted models (Gemma) with a local TTS model for speech, with no external cloud APIs.
On-device agent orchestration: useful for privacy-sensitive evaluation in research, and for cost and latency in production.
Scope
- Agent orchestration for content generation, scoring, and feedback
- Local Gemma for generation and evaluation
- Local TTS for spoken prompts and responses
Stack
Multi-agentGemmaLocal TTSPythonOn-device