Varsity Interviews

Make that career jump from JV to Varsity with Empathic AI.

Comment

Inspiration

Varsity Interviews was inspired by our own experiences as job seekers. We wanted to create a comprehensive and effective way to help ourselves and others prepare for interviews. We noticed that many candidates, even those with the right skills, often struggle to articulate their experiences and showcase their abilities confidently. By leveraging AI tools like large language models and sentiment analysis, we decided to develop an intelligent assistant that provides personalized, real-time feedback to help candidates refine their interview skills.

What it does

Varsity Interviews is an AI-powered interview prep platform designed to help job seekers gain a sense of confidence for their software engineering interviews. It conducts mock interviews, provides real-time feedback, and offers personalized tips to improve both technical and behavioral responses. It uses sentiment analysis to help interviewees hone in on their presentation skills and confidence. The app also assesses key aspects of technical knowledge, problem-solving skills, and communication, finally outputting the most strongly conveyed emotions of their responses throughout the duration of their interview. It guides users to polish their performance and boost their chances of success.

How we built it

We built Varsity Interviews using the supported APIs and WebSockets for Hume.ai’s EVI, sentiment analysis, live speech-to-speech, and openAI’s LLM, supported by a framework of React and Typescript markup language. The platform's goal is to provide an AI assistant capable of conducting interviews and analyzing responses, ultimately outputting a summary of the interviewee’s performance. We provided a thorough system prompt while setting up our Hume configuration, and then guided the AI to lead the user through the interview process. We wanted the user interface to be intuitive and user-friendly, so we designed a simplistic yet descriptive UI that sufficiently grants the user functionality that will allow them to make full use of our product.

Challenges we ran into

One of the main challenges we faced was integrating the React front end to the Typescript backend. This was challenging because our team had no experience with Typescript before this hackathon, and the fact that we had to depend on it as the backbone of our solution seemed daunting at first. However, talking to the experts at Hume.ai—Zach, Morgan, and Dylan—we learned as much as we could to take a bottom-up approach and build our solution. Another big challenge was creating a real-time transcript. Figuring out how to access and leverage the configs and chat for EVI using Typescript was instrumental in providing speech-to-text to the users.

Accomplishments that we're proud of

For a team that had half its members enter Saturday morning with 0 hackathon experience, we are proud that each person was able to contribute and learn from the development process. Only 1 of us had ever touched frontend before, and seeing the team grow and learn from each other was a really wholesome experience. Apart from our personal and swe knowledge increases, we’re glad this project can be useful to many students and professionals struggling to land SWE roles in tech companies at the moment.

What we learned

Throughout the development process, we learned the importance of understanding documentation and planning out our design over getting chatGPT to code everything for us. Copying one GPT response could cause cascading errors down the line if we didn’t understand it and confuse us and the mentors when we asked for help, revealing a double-edged sword. We also learned that mentors who have worked on things for years on end are extremely valuable resources and reaching out to them with non-frivolous questions can help facilitate the knowledge transfer from the professionals of today to the rising ones of tomorrow.

What's next for Varsity Interviews

Looking ahead, we plan to expand Varsity Interviews by incorporating more advanced features, such as video interview analysis and role-specific question sets. We also aim to enhance the AI's empathy and adaptability, making it even more effective in providing personalized coaching. Additionally, we are exploring partnerships with educational institutions and career services to offer our platform to a wider audience, helping more job seekers succeed in their career journeys.

Built With

Share this project:

Updates