◇Cheng Chen / Projects
Prototype · AI interview practice

Practice the interview, not just the answers.

OfferPilot builds a structured technical-interview session around your target role and level. It follows up on your answers, evaluates individual questions, and assembles a report you can use for the next round.

View source & setup ↗Explore the flow ↓
Illustrative session · backend not running here
InterviewerHow would you keep an API responsive when a downstream service slows down?
CandidateI would set timeouts and use a circuit breaker to prevent cascading failures.
Follow-upWhat signals would tell you the circuit should close again?
Project previewThis page introduces the open-source app. Live AI interviews require the Next.js server and a user-configured model API; GitHub Pages cannot run those server routes.
How it works

A complete practice loop

The project joins company-style prompts, a multi-turn interview flow, per-question evaluation, and a final report.

01 / CONFIGURE

Set the context

Choose a company, target role, seniority, interview round, and optional résumé context.

02 / RESPOND

Work through follow-ups

Answer technical questions while the interview state tracks the conversation and its next phase.

03 / REVIEW

See the report

Inspect question-level feedback, an overall breakdown, and a locally saved session history.

Run the full app

Bring your own model, keep control of the session.

The full product is a Next.js application, not a static HTML demo. Its chat, evaluation, and reporting routes call the model selected in Settings. Follow the repository README to run it locally.

Open setup guide ↗
  • ApplicationNext.js, TypeScript, Tailwind CSS, Zustand
  • AI integrationVercel AI SDK; provider and API key configured by the user
  • OutputQuestion feedback, report, PDF export, and local history
  • Deployment boundaryServer routes need a compatible host; no API key is embedded in this page.