PROJECT 01 / FULL-STACK / APPLIED AI
Mock Mentor.
A better space to prepare.
An AI interview platform that turns practice into a conversation. Adaptive questions. Isolated sessions. Built for the next attempt.
LET’S THINK THROUGH THIS.
How would you design a system
that scales with its users?
01 / CONTEXT
The problem.
Interview practice needs structure without becoming a fixed script. The challenge is to support a coherent conversation while keeping each interview session independent and avoiding repeated questions.
02 / APPROACH
From question to system.
A full-stack application connects the interview experience to REST API routes and adaptive AI questioning. Session isolation and question deduplication make the conversation flow explicit responsibilities of the backend.
03 / ARCHITECTURE & WORKFLOW
How the pieces connect.
A conceptual overview of the project’s main components.
- 01Interview session
- 02Session context
- 03Adaptive question
- 04Response & follow-up
04 / IMPLEMENTATION
Decisions that matter.
Keep session boundaries explicit
Session isolation prevents the context of one practice interview from becoming the context of another.
Treat repetition as a system concern
Question deduplication sits alongside adaptive questioning so the interview can cover new ground.
Package the runtime
Docker provides a reproducible deployment unit for the application.
05 / OUTCOMES & LEARNINGS
What the work taught me.
The project brings together 15+ API routes, adaptive interview questioning, and a containerized deployment. The central engineering lesson is that a useful AI experience depends as much on context boundaries and application state as on generation.
Built with.
Interface visuals are illustrative, not application screenshots. Explore the repository for source code and the current implementation.
NEXT CONNECTION / 02