Building Mock AI

Mock AI won the SDSU x AWS Ai hackathon! It is a interview simulation platform that provides comprehensive feedback on candidates' performance across multiple dimensions. The core idea was to create a realistic interview environment that could evaluate not just answers, but also tone, body language, and facial expressions.

We used Next & TypeScript, Fastapi, base 64, Amazon bedrock, dynamo, lambda, & transcribe!

The biggest challenge was creating realistic interview scenarios that could adapt to different industries and roles. We implemented dynamic question generation based on job descriptions, resume, and company culture, ensuring that each mock interview felt authentic and relevant to the candidate's target position.

Finally, we faced the challenge of providing actionable feedback that candidates could actually use to improve. Rather than just scoring performance, we developed detailed analysis systems that provide specific recommendations for improvement, including practice exercises and targeted coaching suggestions.