Born from a research problem

In 2008, ECAD, Brazil's central office for music copyright collection, needed to know which song was playing on every radio station in the country, and no vendor could do it. The project was contracted as applied research with PUC-Rio. InnoVox's two founders, then PhD candidates, took on its technical core: automatic audio identification.

InnoVox was founded to take that research into production. The relationship with ECAD is still active today, and the solution is now in its fourth generation.

Since then we have followed the same path on other problems with no ready-made solution: from problem to research, from research to engineering, from engineering to production.

Track record

Before and throughout InnoVox, the founders have worked on signal problems where reliability matters more than demos:

automatic identification of advertising and music on radio and TV, for Crowley Broadcast;
strategic communications, secure voice communication equipment for the Amazon, and automatic speaker recognition for the Brazilian Army;
software-defined radio for the Brazilian Army;
wireless time and frequency transfer for INMETRO, Brazil's national metrology institute;
forensic phonetics (voice comparison and speaker identification) for public security agencies.

Institutions and companies we have worked with:

ECAD
PUC-Rio
IME
CTEx
Crowley
Acceta

What guides us

Problem before technology. We start by understanding what needs to be solved.
Research when it's needed. When we don't know, we investigate.
Technology that works. Research has to survive the real world.
Trust. We don't promise what we can't deliver.
Ethics. Not every problem that can be solved should be solved by us. We don't build technology whose business model depends on exploiting or manipulating human vulnerabilities.

Small on purpose

We are a small, specialized team, by choice. We grow in knowledge, reputation and intellectual property, not in headcount. It means the people doing the research are the people talking to you.

Who we are

César Augusto Medina Sotomayor

Co-founder and CEO

For 20 years he has worked on technology problems that had no ready-made solution, in environments where reliability matters more than demos. What interests him most is the gap between a promising model and a system that holds up in real operation: where AI adds value, where it adds risk, and where the answer must be deterministic and auditable.

B.Sc. in Electronic Engineering from Escuela Politécnica del Ejército (Ecuador); M.Sc. in Electrical Engineering from the Military Institute of Engineering (IME), with a thesis on speech enhancement for automatic speaker verification; and Ph.D. in Electrical Engineering from PUC-Rio (2009), on transmission systems. Postdoctoral research at PUC-Rio and IME (2009–2017) in telecommunications and digital signal processing.

Dirceu Silva

Co-founder and CTO

B.Sc. in Telecommunications Engineering and M.Sc. from the Military Institute of Engineering (IME). He served 25 years as an officer in the Brazilian Army, working on strategic communications projects and on the development of encrypted voice communication equipment for the Amazon, and supervised graduate students at IME in speech signal processing. Specialist in digital signal processing and machine learning, secure voice communication, and automatic speech, speaker and music recognition.

Our own technology

Much of what we build is our own technology, written from scratch: