MockMade

logo
landing page
behavioral interview report
technical interview page

Inspiration

As computer science students ourselves, we experienced the rigorous internship hunt season in this job market. The influx of software engineers led companies to increase their hiring bar for interviews. We wanted to help our peers land their dream jobs by using real-time human-like AI to conduct job interviews, simulating a real interview. We hope that this product can be a helpful practice tool that improves job seekers' interview skills.

What it does

MockMade is a mock interview platform that conducts technical and behavioral interviews in real time. Choose a specific company to practice for and choose either a behavioral or technical interview. Our behavioral interview is conducted similar to a phone screen where the user would get a call and an AI would ask questions pertaining to the job. Once the user completes their interview, we process the audio and transcript to grade their interview based on various categories. While it was a phone call interview, you can watch a simulation of your interview in video form between the AI interviewer and your responses. For our technical interview, you would use a leetcode-like platform integrated with AI. While you're coding your solution, ask the interviewer for tips on how to complete the problem and keep you on the right track till you pass all the test cases instead of just searching up the solution.

How we built it

Our behavioral interview uses Retell AI to power the real-time voice and call functionality. The post-interview analysis is conducted using multi-prompt OpenAI calls to grade each category. Our interview video simulation is powered by Sync, where we upload videos to the cloud using Cloudinary and lip sync them to the interview audios. The technical interview involves Whisper to transcribe the input audio, Cartesia Sonic to generate human-like text-to-speech audio, and OpenAI GPT 4o to parse the input transcript and code. We use Auth0 to secure our website with a robust custom log-in and sign-up page. To store the technical interview leetcode questions, we used MongoDB's noSQL database. We hosted our website and product completely on the cloud using Render for the backend and Vercel for the frontend. To build the frontend, we used Next.js, React.js, Typescript, and MagicUI.

Challenges we ran into

We had some trouble with Retell AI and making sure that the agent was reliably asking meaningful questions to extensively quiz the interviewee. While we were developing the interview video simulation, we had some trouble getting the latency as low as possible and hosting the videos on the cloud effectively. We spent time working to connect all these technologies to work seamlessly with each other to provide a great experience for the user.

Accomplishments that we're proud of

We are proud of creating a solution to a problem that every member of this hackathon faces. We are particularly proud of integrating a phone call to the inputted phone number and interview analysis. The integration with Sync was impressive to see since we were able to simulate the interview in video form. Lastly, we are proud of our integrated leetcode environment, which seamlessly guides a coder through a problem.

What we learned

Throughout this project, we learned how to use and integrate countless technologies we would not have been exposed to before, including Retell AI, Sync, OpenAI, Cartesia, Auth0, MongoDB, Render, Vercel, etc. We learned how to make user-facing designs for our frontend and use clean UI components to represent our product in the best way.

What's next for MockMade

We're looking to go public with our product and have real-world impacts on job seekers. We're looking to expand our services to all top companies by using the most frequently asked interview questions. Our phone call agent will be more robust and ask more insightful follow-ups. We plan to use a variety of leetcode questions for the technical interview portion to target more company-specific topics.