Deloitte · AI Intern

MAY TO JUN 2026 · BENGALURU

Took a research lip-sync model and got it running on live video.

• Rebuilt a research lip-sync prototype into a real-time PyTorch service holding 30 fps on live video.
• Shipped a multi-tenant dashboard and WebRTC streaming layer serving 3 enterprise clients under 200 ms latency.
• Containerized the service in Docker with deployment docs consultants used to run demos unassisted.

PyTorch, MuseTalk, WebRTC, Docker