SDE Intern @ Soundverse AI
- 1.1
Scaled a Whisper transcription and speaker-diarization pipeline from 25 MB to 500 MB inputs with silence-aware chunking, parallel transcription and live progress over Server-Sent Events.
- 1.2
Built core services of the Music Distribution pipeline that releases artists' songs to Spotify, Apple Music and YouTube, behind a fail-closed quality gate of 7 weighted checks (loudness, ISRC conflicts, AI provenance, copyright, AWS Rekognition moderation).
- 1.3
Built the Media Converter end to end for audio, video and images: background jobs with live Socket.IO progress, per-plan Redis rate limits and token billing; fixed a proxy that stalled large files for 40–200 s.
- 1.4
Co-built the move from MP4 to HLS over a CDN, benchmarking seeks 7.6x faster and first byte 6x faster.
- 1.5
Designed a cross-channel notification system where one record per event feeds the in-app bell, Firebase push and Amazon SES email, cancelling held emails once the user reads it anywhere.
- 1.6
Worked across gRPC/Protobuf and REST API design, PostgreSQL schemas and integrations like Didit identity verification, Sign in with Apple and Google Play Billing.

