Intraviewer

Inspiration

Preparing for a job interview is stressful, especially for people early in their careers who may not have much interviewing experience. Many people do struggle when asked questions that represent their experiences and general questions about themselves. While there are lots of online resources that tell you what to do during an interview, we wanted to create something that makes the interview prep process more personalized and interactive. People shouldn't need a career counselor or coach to get a job. This inspired us to build Intrviewer, an AI-powered career prep platform that helps users build their confidence and improve how they present themselves professionally.

What it does

Intraviewer prepares you for job interviews. You can enter your job title, experience level, and job description to generate a few interview questions that are fully personalized. The platform has a Mock Interview feature and a Guided Interview feature. The Mock Interview has 5 questions and gives feedback at the end, and the Guided Interview gives feedback after every question and gives detailed feedback on how to respond to the question better. There's also a resume review feature that allows you to upload your resume, and it takes into account the questions in the mock/guided interview to make the website more personalized.

How we built it

We developed Intraviewer using HTML, CSS, and JavaScript for the frontend, along with Python and Flask for the backend. We integrated Google's Gemini API to generate personalized interview questions, analyze user responses, and provide constructive feedback. We also used the pypdf library to extract information from uploaded résumés, allowing the AI to tailor interview questions to each user's background. ElevenLabs was also used to create the voice of the AI speaker who acts as a hiring manager to make the mock interview more realistic. Our team used Visual Studio Code, Live Share, and GitHub to develop and organize the project.

Challenges we ran into

Our biggest early challenge was that communication between Flask and the Gemini API didn't align. Another challenge was that, while working in the same workspace with the team, we ran into an unexpected bug where Codex wouldn't connect to the files. We also had to figure out how to extract text from the resume accurately and make sure the AI generated feedback correctly. We debugged and communicated through every mistake and problem we came across.

Accomplishments that we're proud of

Our biggest accomplishment was turning our first real idea into a fully functional website within the time of a hackathon. We successfully integrated Google's Gemini AI to generate personalized interview questions that provide feedback based on users' responses and professional backgrounds. We're also proud of creating multiple career preparation tools within one platform, including résumé reviews, elevator pitch coaching, and interview simulations. Being able to combine these features into a single accessible website was especially rewarding because it demonstrated how AI can be used to address challenges that students and job seekers face.

What we learned

Throughout the project, we learned how to build a full-stack website and connect multiple programming languages to create an interactive user experience. We gained experience working with Flask, JavaScript, CSS, REST APIs, and Google's Gemini AI. We also learned how important prompt engineering is when developing AI-powered applications, especially when generating personalized questions and useful feedback. This hackathon also helped us improve our teamwork, time management, and problem-solving skills. Working under a tight deadline taught us how to prioritize important features, troubleshoot errors efficiently, and adapt when things didn't go as planned.

What's next for Intraviewer

Moving forward, we plan to expand Intraviewer from a hackathon prototype into a fully robust platform for career readiness. Our upcoming roadmap includes integrating real-time speech-to-text and voice analysis to help users practice their answers aloud, adding expanded role specializations for technical, product, and financial fields, and implementing user accounts to track progress over time. We also hope to build features that allow users to save and download their recorded mock interview sessions with peers or mentors for collaborative feedback.