PayTunes
Co-Founder & CTO · 2021–present · paytunes.in
The Problem
India had no programmatic audio advertising infrastructure. Radio was analog, podcasts were nascent, and digital audio inventory had no scalable way to reach advertisers in real time. The market needed a full-stack DSP — not just a media buying tool, but the underlying bidding engine, delivery network, and analytics platform from scratch.
The System
The core is a high-performance Real-Time Bidding engine purpose-built for extreme concurrency and ultra-low latency. End-to-end bidding lifecycle completes under 100ms.
scale: 350M+ daily requests · 2B+ weekly · 3,000–3,500 RPS sustained
latency: sub-100ms end-to-end bidding lifecycle
regions: Hong Kong · Singapore · Frankfurt
stack: Dockerised stateless services · Aerospike · Redis · MongoDB
cost: ~$10K/month infrastructure · 40% cost reduction via localised caching
Beyond the RTB engine: full campaign management, tracking, analytics pipelines, reporting, and multi-region media delivery. Proprietary IP includes Dynamic Creative Optimization (DCO), hyper-local targeting, the VOXALIS AI TTS engine, and sequential retargeting.
The Scale
400M+ unique listeners through 850+ publisher integrations including Spotify and Delhi Metro. 1.4B+ annual impressions. 1,000+ brand clients including Amazon Prime, ICICI Prudential, KFC, and TATA AIG.
VOXALIS — India's first AI metro infotainment system — runs on PayTunes infrastructure. Context-aware audio across 6+ cities, reaching 300,000+ daily commuters.
Category leader: Tracxn Top Audio AdTech, Editor's Pick 2024–2025. ExchangeWire Global Awards 2024 shortlist for "Best Regional Campaign — APAC" with Triton Digital.
Philosophy
Complexity is the enemy of scale. The entire system runs on ~$10K/month because every architectural decision optimised for simplicity at the edge — stateless services, aggressive caching, single-region compute with multi-region data. The HackerNoon article below is the full technical write-up.