Computing
Public sourceCaptureSuite
Multimodal recording means several devices, each with its own software, and a session that has to be stitched together afterward. I lead the architecture and direction of CaptureSuite, which is meant to be one place to run that kind of session and come out with a single sealed package.
A long-running daemon keeps the clock and the session state. Each device lives in a separate worker plugin, written in C++ or Python, so adding a new sensor doesn't mean touching the core. There's a desktop interface for the operator, a simulator so it can be developed without hardware, and analysis jobs for quality checks and features that read the sealed packages.
If you want to poke at it, the quality-check engine runs on a small synthetic session with no hardware at all.
Windows-only for now. The repository is the place for setup, licensing and current limits.