Research · Stony Brook University · Drive Assist
Drive Assist Activity Editor
A Scratch-style, block-based editor that lets clinicians build interactive research activities for the mindLAMP platform without writing code.
- Role
- Software Developer
- Timeline
- January 2026 – Present
- Stack
- ReactNode.jsRedisDocker
Drive Assist is a driving research study with neurodivergent participants. Its activities run on mindLAMP, a digital mental health platform. Collaborating alongside a team at Harvard, I designed and built a game engine and activity editor from scratch so clinicians can author and scale their own interactive activities across mindLAMP.
The editor has a few parts. In the asset editor, clinicians create the pieces an activity uses, like a traffic light with green, yellow and red states, drawn from simple shapes. The behavior editor works a lot like Scratch: they define the activity's flow by snapping blocks together, like "show message … and wait", "repeat … times", if/else branches, variables, and more. For scenario-based activities, a timeline editor lets them load assets into a scene, animate and move them second by second, and mark which answer is correct. A live preview replays the activity after every change.
I also build custom React activities that use device telemetry to collect real-time behavioral data, and I work on the Node.js, Redis and Docker backend that serves them.
gallery / 5 items

In the activity
What a participant sees while playing Stop Time: a driving situation plays out, then they choose Continue, Slow Down, or Stop. The bar at the top tracks their progress and correct answers.