Learning While Building
MVPs, Prototypes, and the Importance of Physical Gesture
The “build, measure, learn” loop and the importance of iterating on “working software” brought to us by the eXtreme Programming (XP) and Agile Software Development movements is now an essential part of 21st century software development. The ability to iterate, learn, and pivot becomes even more valuable in fast moving industries like digital audio. We can increase our pace of learning and iteration by not only embracing these techniques from XP and Agile by bringing the learning process earlier before you begin to write “working software.”
Will people even want what I’m building? A discussion of Lean Software Development will illuminate ways a Minimal Viable Product (MVP) can be built to get external feedback from the market while writing little—or hopefully no—code. After covering general examples of MVPs, we will look at some fantasy audio plug-in MVP examples as well discuss how MVPs were used to test some currently shipping audio plug-ins.
There is also internal learnings to be gained about the thing we are building: enter the world of prototyping. We will discuss the differences between MVPs and prototypes, how to use both for learning, and about different types of software and non-software prototyping. After a tour of example prototypes, we will look at some prototyping tools useful for digital audio: starting from one of the oldest tools on the block Max (or now Max4Live) to some of the latest community built tools for prototyping machine learning within the DAW.
This tour of user research, MVPs, and prototyping will show many ways to learn along your software development process, but at the end of our journey we will see that there are learnings about an idea that can only be obtained by combining physical gesture with audio feedback.
Roth Michaels
Director of Software Research
inMusic
Roth Michaels is Director of Software Research at inMusic, the family of music technology and consumer electronics brands behind Akai Professional, Moog Music, M-Audio, Denon DJ, and Native Instruments, among others. He stepped into this role following inMusic's 2026 acquisition of Native Instruments, taking on responsibility for the AI/ML research team, leading adoption and training of internal AI tools across the organization, and coordinating software engineering topics that cross brand lines.
Before the acquisition, Roth was a Principal Software Engineer at Native Instruments, where he worked on software architecture to bring together the company's merged engineering organizations and legacy codebases — spanning Native Instruments, iZotope, and Brainworx during the years those brands operated under one roof. He also supported the Audio Research team, helping accelerate the path from research to productization and building fast prototyping tools for product teams. Earlier, as the lead library designer at iZotope, Roth built a new internal cross-platform library, "Glass," part of which is now available as open source. In his prior role as Mix/Master Software Architect, he led development of the reference implementation that moved iZotope's products to subscription and headed the team that launched the company's first SaaS offering for music producers.
Roth studied music composition at Brandeis University and continued his studies in the Dartmouth Digital Musics program. He began his career writing software for his own compositions and for the works of other composers and artists, and teaching MaxMSP to composers and musicians through both private instruction and university course design. Before joining iZotope, he worked as a consultant for small startups building mobile applications specializing in location services and Bluetooth.