Interview Trainer — Mattia Smacchia

The problem

Interview preparation is generic. You rarely know which questions a specific role will bring, or how good your answers are.

The solution

Paste a job posting, or just a job title, and get the ten questions you are most likely to be asked. Draft an answer to each and the app rates it from 1 to 5 with suggestions to improve it.

An API and a chat-style frontend. FastAPI calls the models through OpenRouter, and a React app shows the questions and ratings, with the cost of every call.
Strict JSON output. Both model calls must return a fixed schema, which keeps replies predictable and doubles as a guardrail against prompt injection.

Demo

From a job posting to ten likely questions. Two answers are written and sent for evaluation, and each comes back with a 1 to 5 rating and suggested improvements.

The problem

Interview preparation is generic. You rarely know which questions a specific role will bring, or how good your answers are.

The solution

Paste a job posting, or just a job title, and get the ten questions you are most likely to be asked. Draft an answer to each and the app rates it from 1 to 5 with suggestions to improve it.

An API and a chat-style frontend. FastAPI calls the models through OpenRouter, and a React app shows the questions and ratings, with the cost of every call.
Strict JSON output. Both model calls must return a fixed schema, which keeps replies predictable and doubles as a guardrail against prompt injection.

Demo

From a job posting to ten likely questions. Two answers are written and sent for evaluation, and each comes back with a 1 to 5 rating and suggested improvements.