Built by students who needed better reps.
Mocki is a team-built AI mock interview platform from BeaverHacks 2026. It was made for candidates who want practice that reacts to their resume, target role, and actual answers instead of walking through another static list of questions.
BeaverHacks 2026
NVIDIA track winner.
The product uses a multi-agent interview flow powered by NVIDIA Nemotron, with supporting voice and speech systems to make interview practice feel closer to a real panel.
Resume-aware
Panel-style
Debrief-driven
The team
The builders behind Mocki.
Mocki was built as a team project, with each person helping turn interview prep into something more repeatable, realistic, and useful.
Built with
A practical stack for a real interview loop.
NVIDIA Nemotron Nano 9B
NVIDIA Magpie TTS
NVIDIA Riva ASR
TanStack Start
Supabase
Vercel
Railway
Feedback
Help shape the next version.
Tell us what worked, what broke, or what would make Mocki more useful before a real interview.