Interview Prep Platform
An AI-powered interview preparation platform that identifies knowledge gaps, generates personalized learning plans, and supports AI-powered mock interviews.
JavaPythonLLMsVoice AIReactAPIs
Designed, built, and operated by Amal Chaitanya
Problem
Generic LeetCode grinding does not surface actual gaps. Candidates rehearse strengths while system design and behavioral depth stay weak.
Why I built it
I wanted a loop — diagnose, teach, test by voice, re-diagnose — that adapts to Java/backend and system-design roles instead of one-size-fits-all question banks.
How it works
- Diagnostic quiz + resume skills map to a gap profile.
- Planner generates a week-by-week plan mixing concepts, coding, and design.
- Voice interviewer runs timed mocks and transcribes + scores clarity, correctness, and depth.
- Feedback links each miss back to a lesson, closing the loop.
Technology stack
- Next.js frontend
- Java/Spring Boot + Python services
- LLM orchestration for planning and scoring
- Speech-to-text + TTS for voice mocks
Challenges
- Scoring open-ended system design fairly
- Keeping voice latency conversational
- Avoiding leaky memorization of popular questions
What I learned
- Personalization beats volume — 20 targeted reps beat 200 random ones.
- Voice reveals gaps text hides: hedging, rambling, missing tradeoffs.