Finalist · ATOM × AIS Technology Innovation Challenge
TalentFlow AI
Team of four · ATOM × AIS Challenge, sponsored by EY · Mar 2026
An internal talent mobility platform that turns résumés, video and certifications into structured skills, then matches employees to internal roles with a gap analysis and a learning path.
1What it does
Most companies know less about their employees' skills than LinkedIn does. TalentFlow AI converts the experience data a company already has into structured skill intelligence, and uses it to fill roles from inside.
The pipeline has four stages: Extract → Match → Explain → Grow. Speech-to-text handles video, a language model extracts and scores skills, a matching engine compares employees to open roles, and each match comes with an explanation, a gap analysis and a personalised learning path.
2Result
The project was a finalist at the ATOM × AIS Technology Innovation Challenge at James Madison University.
Stack: Next.js · Supabase · Whisper · LLM extraction