MuseRoom

I worked across the workstation and its AI assistant, connecting a musician’s request to the session’s controls.

Software Engineer & Agentic AI Developer
React / TypeScript · Cloud services · Native audio integration

MuseRoom campaign artwork illustrating a multitrack music session and its Mixwell assistant — MuseRoom campaign illustration · The workstation view is a product visualization.

Working inside a music session

When a musician is shaping a track, each adjustment belongs to an ongoing thought: move a clip, listen again, bring an instrument into focus. MuseRoom brings an AI co-producer into that process, alongside the timeline, piano roll, mixer, and familiar transport controls.

As a software engineer and agentic AI developer, I worked across the browser experience, cloud services, and native audio integration. My contribution sat at the connection between what the musician asked for, what the session understood, and what the workstation actually changed. I helped develop that experience within a shared product and codebase.

Connecting the workstation

MuseRoom spans systems with different responsibilities. The React and TypeScript interface presents tracks, clips, and controls. Cloud services support project storage and assistant workflows. A native JUCE/C++ audio layer handles playback and plugin processing in the desktop application.

Those layers need to describe the same session. A change to a track should affect the sound as well as the control on screen. A play command should follow the same transport behavior whether it comes from a button, a keyboard shortcut, or the assistant. That made integration part of the musician’s experience, even when the work itself was below the interface.

I worked on session interaction and assistant behavior with that continuity in mind. My changes included command planning, multi-track solo behavior, and opening and sorting projects through Mixwell. These were concrete points where an apparently simple request depended on the application’s existing state and controls.

A musician’s request

Mixwell & the session controlsInterpret the action · Identify the track · Update session state

Browser audioReact controls & Web Audio

Native audioDesktop bridge & JUCE

Architecture illustration · The interface and the audio path must describe the same musical session.

Connecting Mixwell to the controls

Mixwell makes musical language an entry point into the workstation. The assistant work explored requests around playback, tempo, loops, track volume, and project navigation. Turning those requests into useful actions meant identifying the right track, interpreting the requested change, and connecting it to the session’s controls.

I also worked on agentic tool workflows and the cloud connections around the product. The broader engineering challenge was keeping conversational assistance connected to the tools a musician was already using.

MuseRoom developed my practice across interface engineering, audio systems, and AI interaction. It gave me a concrete way to think about creative software: the visible control, the underlying state, and the resulting sound all contribute to the same moment of making music.