Josh Greifer
Chief Scientist
Engineer and musician who helped transform Steinberg Cubase from an Atari ST sequencer into a full recording studio for Mac/PC — establishing a new standard for low‑latency audio. Since then he has worked across finance, biomedical signal processing, media infrastructure, and speech‑AI productisation.
Selected milestones
- University of Sussex (1977–92): Studied Philosophy with cognitive science, including Natural Language Processing and AI with luminaries such as early contributor to natural language understanding, Gerald Gazdar.
- Steinberg Media Technologies (1988–92): ported Cubase core engine to PC; developed Cubase Mac; architected low‑latency audio pipeline. (Charlie Steinberg was still hands‑on — occasionally sleeping on Josh’s sofa during sprints.)
- Goldman Sachs (1994–2004): technology development spanning risk systems, portfolio analysis, and client margin.
- ToHealth — formerly HealthSmart (2005–14): R&D and signal processing; ran a full clinical trial for non‑invasive atrial fibrillation diagnosis using Poincaré plots of pulse signals from smartphones and other sensors.
- Speech Engineering Ltd (2014– ): Chief Architect for Eloqute, a patented continuous‑speech pronunciation assessment & remediation system.
- BBC (2014–19): contractor within BBC R&D and Platform Technology (including BBC Redux).
- Neurence (2019– ): Senior ML Engineer; team reduced end‑to‑end GAN audio processing latency from >250 ms to <50 ms.
Now: BabelBit
Joining to drive latency‑critical audio pipelines and streaming inference for the speech‑to‑speech engine, including phonetic prediction and an evaluation harness that balances latency with accuracy.