Introducing: Prosody
Inspiration
Preparing for interviews can be a daunting task, especially when it comes to practicing soft skills and technical knowledge. Prosody is designed to help bridge this gap by generating an AI-powered, adaptable, and fully automated interviewer to provide a seamless and immersive interview experience.
What it does
Prosody utilizes state-of-the-art technologies such as GPT-4, AWS, Python, React, and Hume, to provide a user-friendly platform that offers a series of engaging and personalized interviews. The LLM is equipped to hold cohesive, back-and-forth conversations that adapt to the content and emotions displayed by the user. The platform caters to both technical and non-technical interviews, with the ability to interpret code and provide helpful feedback.
Key Features:
- Adaptive AI Conversation: Using GPT-4 and Hume, the AI interviewer adjusts its questions and feedback based on user responses and emotions detected.
- Real-time Weakness Detection: The generative AI, GPT-4, efficiently identifies weak points of the interviewee's performance and pinpoints where exactly in the interview they occurred.
- Powerful Analytics Dashboard: Gain insight into your performance with comprehensive analytics that reveal patterns and areas for improvement.
- Code Interpretation & Feedback: For technical interviews, the AI Interview Master has the capability to understand code, suggest optimizations and provide accurate feedback.
- Multimodal Applications: This project can be easily expanded to cover other contexts such as tutoring, debate coaching, and test preparation by specifying the required context for GPT-4.
How we built it
We combined cutting-edge technologies in a seamless manner to build a comprehensive and powerful solution.
- GPT-4: The core AI engine, responsible for generating conversation, providing feedback, and adapting to the user's input and emotions.
- Hume: Utilized for emotion analysis, enabling the AI to adjust its responses and connect with the user on a deeper level.
- AWS: Our cloud platform for hosting, storage, and compute power.
- Python: Backend application programming using Flask and RESTful APIs.
- React: Robust frontend development for a responsive and intuitive user interface.
Challenges we ran into
Integrating the various technologies and ensuring they work seamlessly with each other was a major challenge. Additionally, finding the right balance between user privacy and generating valuable feedback was crucial to creating an effective AI interview experience. Ensuring that the AI could accurately interpret code and provide meaningful suggestions was also a significant hurdle. Moreover, towards the end of the project, as we were about to deploy, we ran into a multitude of errors regarding file-formatting and other such issues that severely impeded our workflow/momentum.
Accomplishments that we're proud of
We are proud of successfully integrating multiple technologies into an innovative, user-friendly, and effective platform for interview preparation. This platform not only provides a personalized interview experience but also helps users receive constructive feedback, highlights their weaknesses, and offers valuable insights for improvement. We also demonstrated the multimodal capabilities and scalability of the project in various other contexts, such as tutoring, debate coaching, and test preparation. Moreover, our team consisted solely of rising freshmen, with minimal industry, or even hackathon experience. We've loved being able to explore meaningful applications of AI.
What we learned
Our journey taught us the importance of integrating various technologies while building a scalable and robust AI-powered solution. We learned how to leverage GPT-4 effectively for generating cohesive and context-aware conversations, and we explored the capabilities of Hume in detecting emotions and personalizing responses. Moreover, we developed a greater understanding of using AWS cloud services, Python, and React for creating a seamless user experience.
What's next for Prosody
The future of Prosody involves expanding its capabilities to cater to a broader range of interview types and incorporating additional languages for a more accessible platform. We also envision the ability to connect users with industry experts for personalized feedback along with the AI-generated insights. Furthermore, we plan to develop mobile applications to make the platform even more accessible and convenient for users. This project has immense potential for growth across various domains, as its core functionalities can be repurposed for multiple use-cases such as virtual teaching assistants, personal development coaches, or conflict resolution simulations.
Built With
- Python
- Flask
- React
- AWS
- GPT-4
- Hume
Creators
Built With
- amazon-web-services
- deep-learning
- hume
- llm
- openai
- python
- react
- whisper
Log in or sign up for Devpost to join the conversation.