Illumine

What it does

Illumine automates traditional teaching and evaluation using Gemini AI, leveraging LLM prompting techniques to achieve desired outcomes. Instructors upload course content, which Gemini AI processes to generate a structured curriculum. We utilize Role, Task, and Format (RTF) Structural Prompting to simulate course creators, evaluators, and instructor personas, enabling the AI to effectively organize content into a weekly curriculum with teaching notes and in-class activities.
For evaluations, instructors specify the number and types of questions. We apply Recursive Task-Based Prompts in Zero-Shot and Few-Shot modes to guide the AI in generating assessments, including multiple-choice, fill-in-the-blank, true/false, and short answer questions. These techniques ensure that each question aligns with course content and educational objectives while generating accurate answer keys.
During grading, chain of thought (CoT) prompts are used to instruct Gemini AI in evaluating student responses, determining correctness, and providing concise explanations. This ensures feedback is contextually relevant with no hallucination, and enhances the learning experience.
By integrating these advanced prompting techniques at each step, Illumine maximizes Gemini AI’s capabilities, delivering a tailored, efficient, and scalable educational solution. This approach is especially beneficial in under-resourced settings, where maintaining high educational standards is challenging.