A digital twin for how a brain watches your ad
Engram predicts second-by-second cortical engagement for video creative, then turns that signal into edits you can actually make before the spend goes live.
Who we are
Computer science student at the University of Pennsylvania, with an interest in the human brain for as long as memory serves.
AI Engineering student at the University of Pennsylvania, with machine-learning research experience at Penn's GRASP Lab.
What we're building
Most tools grade a video after the fact, with a single score or a vague sense of "good retention." Engram is built on real neuroscience: a brain-encoding model trained on fMRI recordings of people watching video, adapted from MedARC's Algonauts 2025 work on the Courtois NeuroMod dataset. We use it to predict activation across 1,000 named cortical regions for every second of your video, split across visual, audio, and language signal, so you can see exactly where attention builds and where it breaks.
Why it matters
Creative teams already know when a cut "feels" off, but rarely why, or which second to fix. Grounding that feedback in a real neural response model means the notes you take into the edit, and into the client review, are specific, timestamped, and tied to a mechanism, not a guess.
Where we are
Engram is in private beta. We're onboarding a small group of agencies and brand teams first to sharpen the coaching before opening up more broadly.