DTVA
I can build and secure a complex, production-grade system.
What it is
DTVA is a secure platform that handles audio and video signal processing at scale. It puts a usable interface on top of heavy DSP work, and runs across web and native mobile.
Outcome
- Live production platform
- Android approved
- iOS in submission
Key Features
Audio/video DSP pipeline
Robust processing for real-time signal integrity and manipulation.
Three-tier subscription
Complex entitlement logic integrating securely with payment gateways.
Native apps
Platform-specific clients optimized for hardware acceleration.
Full RLS audit
Row-Level Security keeps each tenant's data isolated.
AI-assisted feedback
Intelligent system analyzing DSP output to provide actionable tuning recommendations in real-time.
Why this architecture
Audio/video pitch analysis is CPU-heavy and needs Python libraries (Librosa, ffmpeg) that don't run in a standard Edge Function, so that workload was split into a dedicated GCP Cloud Run service, sized for the job (2Gi memory, 300s timeout), while everything else stayed in Supabase. The three-tier structure maps directly to how a vocal coach already packages value (content, deeper analysis, personal time) and caps Dan's own 1:1 hours to the top tier only. Tier access is enforced with Row Level Security at the database level, not just hidden in the UI, so a Foundation subscriber can't reach VIP content by calling the API directly. Native mobile (not just a web app) was necessary because reliable microphone and camera access for recording is a native permissions problem on iOS/Android.