Cognitive Strike

What it does

The app aims to revolutionize preparation for various scenarios by using Gemini. We designed it to simulate realistic video interactions. We fine tuned the System Role such that it would give personalized mock sessions by adapting to specific user needs and helping them prepare for interviews.

The app features two dynamic modes: Chat and Video modes. In Chat mode users engage with a personalized Gemini Interviewer and then take personalized quizzes generated on that specific conversation. Video mode simulates a live video call with the user utilizing Google Cloud TTS and webSTT. This model asynchronously uploads the user's video feed and generate dynamic response from Gemini based on what the model sees and what the user speaks.

PS - The only LLM API used here is Gemini. The other API is Google Cloud Text-to-Speech.