Voicescript Technologies

The creation of a proof-of-concept of a real-time, multi-person transcription system to allow immediate written documentation to be generated from spoken interviews.

I was responsible for creating the initial prototype, both hardware and software, working with teams from the London Mayor’s office, Deloitte, The Met and Nottinghamshire Police, Nuance Communications, and an advanced engineering team from London technical pioneers LShift.

The project had exciting challenges, including engineering a solution allowing audio processing nodes to run on AWS.

I created the prototype front end in Max/MSP, using its Java and JavaScript features, and the backend orchestrator was developed in Scala, connected to a next-generation, profile-less speech-to-text engine from Nuance.

The project was a success and saw prototype systems installed at Hendon Police Academy and in live interview rooms at Nottinghamshire Police.

Due to juggling input and processes from stakeholders, partners and the technical team, a tight project management approach was needed. Development was agile, mixed with more traditional methods from other partners.