HireSight

Inspiration

As software engineering students entering a highly competitive job market, we are experiencing firsthand how unprepared many of feel when we finally land a rare oppurtunity of a technical interview. While CS student knows about leetcode for practice, we feel that solving static problems doesn't simulate the real time pressure nor communication of an actual interview. We built this project to bridge the gap by helping students practice not just typing out code solutions, but also simulating a realistic environment to speak clearly, and maintain proffesionalism throughout an interview

What it does

Our project simulates a realistic technical interview experience by combining live coding with real industry questions, voice interaction as well as real time feedback monitoring physical movements and eyecontact with computer vision to prepare users for the entire interview experience.

How we built it

We built the platform using a full stack architecture of React and Next.js interactive interview interface, including live in browswer coding and visual/audio input and outputted live feedback. The backend is powered by FastAPI and WebSockets to support real time communication, Postgres for storing data, and LLM's are implemented to simulate an interviewer and provide feedback with each simulated interview. Computer vision as well as speech recognition was implemented to enable real time interaction and monitoring, alerting users of when they get distracted or go off track during an interview.

Challenges we ran into

A major challenge was building real-time face and eye tracking in the browser to infer user engagement while maintaining low latency and smooth performance. We also faced difficulties synchronizing live speech recognition with streaming AI interviewer responses, ensuring that feedback arrived at the right moments without interrupting the candidate. Solving these problems required careful handling of asynchronous events, streaming data, and state management across the frontend and backend.

Accomplishments that we're proud of

We’re proud of pushing through and overcoming technical challenges we initially struggled with, particularly in real-time computer vision and AI-driven interaction. Most importantly, we worked together to turn those challenges into a product we find personally useful and hope will value to others and help users prepare for technical interviews.

What we learned

Throughout this project, we learned how to pick up new technologies on the fly and design real time systems. We gained experience integrating frontend input and user monitoring with backend streaming to deliver timely AI feedback. The project also strengthened our ability to optimize system performance and intentionally manage delays in real time interactions. Additionally, collaborating under tight deadlines was a valuable learning experience that pushed us to work efficiently as a team.

What's next for HireSight

Next, we plan to expand the platform beyond technical questions by adding behavioral interview scenarios. We also aim to make the AI interviewer more conversational, with richer back and forth dialogue that more closely resembles smooth, natural human conversation. Additionally, we plan to introduce progress tracking so users can measure improvement over time across both technical and communication-focused metrics.