It all began with Tony Robinson (our CTO) pioneering recurrent neural networks in his PhD at the University of Cambridge in the late 80's. After showing that they can make a huge difference in speech recognition he assembled and scaled the Connectionist speech group at the University to build the first real time large vocabulary continuous speech recognition system. During this time Tony started SoftSound, which he successfully scaled and sold to Autonomy in 2000. At Autonomy Tony successfully scaled the speech team before leaving in 2006 to start Cantab Research (Speechmatics today). Based on 20 years of pioneering speech recognition and deep learning, the original aim of the company was to develop ground-breaking speech technologies. His pioneering work was recognised by SpinVox in 2007, who licensed technology and tasked Tony with building the speech team, which he scaled again before the successful acquisition of SpinVox by Nuance at the end of 2009.
Predicting the GPU and cloud-computing revolution, Tony built on his decades of experience of applying neural networks to the problem of speech recognition and started building full cloud-based speech recognition system. Tony added several highly experienced PhD’s to the team and together via a series of high profile TSB grants, they built one of the most accurate systems yet.
Benedikt joined in summer 2014 and helped launch and scale the company to now 24 people. This rapid growth has been funded entirely through cash flow, for which we were recently awarded the Deloitte Fast50 award.