Jamahook

Jamahook started as a greenfield bet in Zürich: no existing product surface, only the problem of helping producers find and match musical loops fast enough to stay in flow. I designed and built the first customer-facing web app and internal admin panel — loop matching and discovery driven by weights extracted from an AI model, not a hand-curated catalogue alone.

That 0→1 phase was fully hands-on: product shape, React front end, NestJS APIs, and the admin workflows the ops team needed to keep the catalogue honest as usage grew.

• Team stand-up after initial success — once the product proved itself in market, I hired and directed the full-stack pod under ELO, with a deliberate split between frontend (React + Storybook) and Node/NestJS integration work.
• NX monorepo — consolidated web, shared libraries, and admin tooling so multiple engineers could ship without fragmenting the codebase.
• OAuth2 login — standard OAuth authentication for producers signing into the loop-matching platform, alongside the Stripe/GDPR checkout flow below.
• Stripe + GDPR/VAT — checkout, tax, and EU compliance baked into product flows for 15,000+ active users, not bolted on after the fact.
• AWS Lambda admin panels + S3 Glacier — internal tooling on serverless compute and cold storage for massive audio archives.
• Commercial modules in the web portal — payment-enabled subscriptions and licensing, content ownership, vouchers, referral, and affiliate-marketer management, all in the product rather than in spreadsheets beside it.

• Desktop agent with offline matching — a local audio file manager that indexes a producer’s own drive, generating audio metadata and signatures so matching works against their library, not only the cloud catalogue.
• DAW plugins — VST/AU/AAX surfaces so matching happens inside the session, where the creative decision is actually made.
• Plugin server ecosystem — built to scale with usage rather than provisioned for a peak, since plugin sessions arrive in bursts.
• B2B annotation & match recommendation API — developed and maintained as a product in its own right, which is what let the engine be embedded elsewhere later.
• Fraunhofer IDMT collaboration — the acoustic / psychoacoustic annotation and analysis algorithm was developed with the institute behind MP3.

As a direct ELO client in Zürich for over 6 years, the product outgrew the first web-only vision: today it ships as a web app, a DAW plugin, and a desktop application, each with enriched features accumulated over that engagement.

Loopcloud acquired Jamahook in 2026; the dedicated engineering team I helped assemble continues to power Loopcloud’s AI sound matching with that engine.

My role shifted naturally with maturity — hands-on through the early build and team formation, then primarily client and business relationship from 2023 onward while the team owned day-to-day delivery.

I bring high-leverage product engineering and absolute operational discipline: shipping weekly, pressure-testing user inputs, and automating delivery pipelines.

Ask how this applies to you →