Ai&U | Devpost

Ai&U
Interview with AI
Feedback after interview

Inspiration

In the hectic moments before a job interview, users may experience a flurry of emotions, from nervousness to self-doubt, as they grapple with the pressure to perform well and make a positive impression on the interviewer. Our inspiration for Ai&U stemmed from our collective experiences with interviews and findings from the user research we conducted, in which users ranging from under 18 to 34 with diverse educational backgrounds and work experiences described their "nervousness," "lack of feedback," and "difficulty answering certain types of question." We noticed that traditional interview preparation methods, such as rehearsing responses or seeking advice from friends and mentors, often lack personalized feedback and fail to simulate the dynamic nature of real interview scenarios. According to the research survey, most users are comfortable interacting with an AI avatar for interview preparation.

Some users wished for a tool that would enact different company’s interview processes (like a boba shop vs tech corporate), analyze body language and take note of the times that they fidget or stutter, questions based on past interviews, and a way to know what interviewers think of their answers.

Recognizing the growing importance of AI technology in various fields, we saw an opportunity to leverage it to create a solution that addresses these shortcomings. By combining AI-driven simulation with personalized feedback and a friendly user experience, we aimed to provide users with a more effective and engaging way to enhance their confidence in interviews and increase their chances of success.

What it does

Ai&U is a website that utilizes AI avatars to prepare users for interviews. Before the live interview with an AI avatar, users input the job description and resume. AI analyzes this information and prepares specific questions for the interview. During the interview, the AI interviewer reacts with expressions based on the user's responses. We used voice-to-text as well as AI avatars to implement these features. After the interview, Ai&U generates a feedback page along with an AI chat where users can receive personalized feedback and ask the AI interviewer questions directly.

How we built it

We used Figma to design the entire website. Then, we used Next.js to build the frontend and Node.js for the backend,

Challenges we ran into

It was difficult to connect the front and back end through API calls. We didn't have much experience with this, so it was quite challenging to make API endpoints.

Accomplishments that we're proud of

We are really proud of the emphasis on the user throughout the design process. Creating the virtual interview avatar was also really cool, and we are proud that the feature was tested properly.

What we learned

Through our user research and development process, we learned the importance of personalized feedback and real-time interaction in interview preparation. We discovered that users face various challenges, including nervousness, difficulty answering certain types of questions, and lack of feedback on their performance. We also learned a lot about team management and combining frontend and backend.

What's next for Ai&U

Based on the survey results, we want to include the user's past interview results as a way to personalize future interview questions. For example, if the user has been weak in questions that account for "leadership and management" skills, future AI interviews may specifically tackle those types of questions. We also want to develop a feature in which users can choose their AI interviewer (Google software recruiter, BobaGuys manager, etc) for a more personalized and targeted mock interview. In addition, we want to record the user's interview with the AI and have it as part of the feedback page where specific phrases will have timestamps that the AI interviewer wants to comment on specifically.