ConvoCoach | Devpost

Initial Screen
File Uploaded
Select Practice Mode
Upload mp3
Results - Analysis of Bad Recording
Results - Analysis of Good Recording

Inspiration

Institutions that lack resources often struggle to offer students opportunities for professional growth. So we created ConvoCoach with the intent of making professional growth more accessible using the power of AI. The inspiration was to blending technology with educational progress to create an interactive way for people to sharpen their professional skills. Whether it's preparing for an interview or delivering a quick elevator pitch, Convocoach offers quick, AI-driven feedback to help you improve your skills!

Our goal was to build something that makes it easier and more accessible to engage in career development: ConvoCoach addresses the need for personalized coaching, especially for those who may not have easy access to professional resources. The site helps users gain confidence by allowing them to practice and refine their communication skills, all while promoting social good by offering this as a free resource to help underserved communities.

ConvoCoach is also about inclusivity, providing a welcoming platform for underrepresented groups, offering a space for them to boost their skills and step into careers or academic opportunities with greater confidence, as it can be daunting to navigate professional environments where you feel like an outsider.

What it does

Convocoach is an AI-powered platform that lets users practice interviews and elevator pitches by uploading their responses. It listens to how you answer, then provides tailored feedback on things like clarity, tone, confidence, and other criteria specific to your desired interview type. It's designed to help you improve both what you're saying and how you're saying it, offering easy-to-understand suggestions to enhance your overall communication.

By creating a space where users can continuously practice and receive helpful insights, Convocoach empowers people to be better prepared for real-world interactions, making communication more effective and accessible for everyone.

How we built it

We developed the frontend using HTML, CSS, and JavaScript, employing a dynamic interface that allows users to upload audio files in real time. We implemented responsive design elements and interactive buttons for selecting between different modes—behavioral interview, technical interview, and elevator pitch. The frontend communicates with the Flask backend through asynchronous API calls.

On the backend, we used Flask and Python to handle the upload and storage of audio files and to make requests to Google’s Gemini AI. The backend processes the audio files, passes them to Gemini, evaluates the audio data against multiple criteria, and returns a detailed analysis based on criteria for each interview mode.

Challenges we ran into

We had to replace the site's AI feedback section with a browser pop-up with the feedback because we could not load the AI responses back onto the website's display.

Accomplishments that we're proud of

We were able to build a functional website which has the ability to upload, read files with AI, and produce tailored feedback.

What we learned

Learned a lot about using Flask for the first time.

What's next for ConvoCoach

We are aiming to implement a key functionality that allows users to record audio directly within the app, convert the recording into an MP3 format, and enable easy export of the file. This feature would streamline the process of creating and sharing audio content, enhancing the app's usability by eliminating the need for external tools. By integrating in-app recording and direct MP3 conversion, users can efficiently generate audio files without leaving the app, providing a smoother and more seamless experience.

In addition to this, we plan to introduce improved UI/UX elements, such as the ability to open and close a modal at will. This modal will serve as a space to view and store any newly generated AI content as well as preexisting creations. These enhancements aim to offer users greater flexibility and control, allowing them to easily access, manage, and interact with AI-generated content within a dynamic and user-friendly interface.