Appu (Apu) Shaji

speaker

101 appearances 1 recordings 1 series first heard Jul 2026 last heard 14 Jul

Appu (Apu) Shaji’s voice in public audio — every appearance, attributed to the second.

Trend

recordings per month · last 12 months

Recordings per month over the last 12 months — 1 in all, peaking in Jul 2026 with 1.

Appearances

newest first · ▶ plays the moment

Working Smarter · Building AI that can search inside videos (and photos and audio too) · 14 Jul 2026 podcast

So it has seen people showing a visual and explaining these things so that when a new scene comes in, it's able to activate similar patterns.

In some cases, we run maybe not one model, multiple models, one model to capture the visual part, one model to capture the audio part, and then combine all the modalities together.

If we can describe it textually, then we can model it.

Okay.

But there are also things which is beyond textual representation, things regarding aesthetics, for example, like some is a long shot, some.

It's a wide shot.

There are a lot of these filmmaking techniques that we are able to capture.

Similarly, for auditory auditor thing, some of the music is in a certain genre.

So, for example, something is very dramatic or something is quite soothing.

So we we are able to go and parse that kind of information also out.

And one of the paradigms, especially in the last four to five years of machine learning research, is that.

People don't hand specify when something happens, do this.

Rather, it tries to infer these patterns internally using a lot of data.

And when a new stream of data comes in, a new video for example comes in, it's able to apply these patterns it has learned in the network into these novel use cases.

So it's an interesting question in the sense that if this end state was presented to me like if you g if if you go back ten years back and said this is what we have achieved, I would have been super happy essentially.

But right now, what we build

six months before is getting deprecated within six months because the community is moving very fast.

Mm-hmm.

So right now I'm much more excited because we feel the scope has actually increased much more drastically.

And again, we still don't have the data of Star Trek, so there is still big things to build essentially.

Showing 81–100 of 101 · page 5 of 6