Tormod Ree is the chief product and engineering officer at Neat, the Zoom-backed video hardware company. Before Neat he co-founded AVA, a computer vision security company Motorola acquired, and spent close to eight years at Cisco running the Spark Board. On Chain of Thought he explained how Neat turns a room of 11 or more cameras and dozens of microphones into a system that directs the meeting, why almost all of that AI has to run at the edge, and how the company builds computer vision models without ever collecting a customer’s audio or video.
Episodes
-
EP 70 · AI Hardware
11 Cameras, Dozens of Mics: The AI That Reads the Room | Tormod Ree, Neat
What the conversations cover
These notes summarize the episodes featuring Tormod Ree, including other speakers in panel conversations. Follow a source link to read the discussion in context.
EP 70 11 Cameras, Dozens of Mics: The AI That Reads the Room | Tormod Ree, Neat
- Tormod Ree says time sensitivity and privacy are the main reasons Neat processes meeting audio and video at the edge. He places cost behind those constraints when deciding where that processing should run. Transcript at 5:22
- Ree explains that Neat cannot simply collect customer meetings as training data. He describes using internally sourced material and relevant public datasets, with relatively little synthetic data at the time of the interview. Transcript at 7:51
- Ree describes distributing inference across devices in a meeting room. A main device coordinates media and metadata from the other devices, reducing the inference work that the central coordinator needs to perform. Transcript at 11:09
In their words
So it's like the harness and the specialized data is, is where we have, have the magic sauce, so to speak.
Tormod Ree 11 Cameras, Dozens of Mics: The AI That Reads the Room | Tormod Ree, Neat · 6:27 But both then sort of the inference and sort of the harness and the orchestration in our case happens at the edge.
Tormod Ree 11 Cameras, Dozens of Mics: The AI That Reads the Room | Tormod Ree, Neat · 11:09 Transcript excerpts; ellipses mark omitted words. Some automatically selected quotes have filler words removed.