Wavelength

Read the room. Even through a screen.

A prototype that uses computer vision during video calls and provides a debrief afterwards.

See the hackathon project

Illustration of three people talking, from the Wavelength landing page — Original artwork from our Wavelength landing page.View full image ↗

MY CONTRIBUTION

Lead computer vision and full-stack developer

HIGHLIGHT

Top 7 + Best UI/UX · MLH × DigitalOcean

TOOLS & METHODS

MediaPipe · React · Node.js · PostgreSQL · DigitalOcean

01 / THE PROBLEM

The problem.

Video calls make it harder to notice the non-verbal cues we rely on in person. We explored whether computer vision could surface useful signals without overwhelming the conversation.

02 / MY CONTRIBUTION

What I worked on.

01
Led the computer vision implementation using MediaPipe and connected live signals to the web experience.
02
Worked across the React frontend, Node backend and deployment to make the concept usable during a real call.
03
Built a post-call debrief flow that turns the session into something a person can reflect on.

03 / PROGRESS & RESULTS

Where it stands.

We finished in the Top 7 and won Best UI/UX Design at the MLH × DigitalOcean hackathon in San Francisco in July 2026.

Hackathon: July 2026