Co Founder @ParjanyaHQ. Developer on platforms. Spatial Computing. Content producer. Married to @raksha_rao. Dad.
Bengaluru, India
Joined March 2009
- Tweets2.5K
- Following299
- Followers961
- Likes1.6K
Pinned Tweet
We did it. Katha Room is nominated for the Apple Design Awards final.
After 16+ years of building for the Apple ecosystem, 6 years of starting and growing Parjanya, and 2 intense, beautiful, challenging years of building Katha Room, we are beyond humbled, thrilled, and honestly still processing this.
Years of staying committed to showcasing India to the world through Prayoga, Vedike and Immersive India, this recognition feels special.
Thank you @AppStore for recognizing this.
One of the most unforgettable moments of our entrepreneurial journey, getting to meet @tim_cook and personally demo our app to him. We will always treasure the moment.
Sad to see him step down as CEO, but incredibly grateful for the impact he’s had on Apple, developers, and dreamers like us.
Excited to see what comes next.
Wanted to learn more about the light injection technique and hence created a demo using Metal and driving it directly from the iPhones front TrueDepth sensor.
Depth-aware light injection in TypeGPU
I got a 448x448 monocular depth model down to ~8 ms on my M4 Pro across ~250 dispatches, which is fast enough to use in realtime :D
Since the inference is written directly in TypeGPU, I can just feed the depth buffer straight into the lighting pass. It never has to leave the GPU or go through any extra synchronization/interop step
Inference, lighting and draw all go through the same command encoder.
What an incredible demo! Relighting has always been a fascinating subject and this demo shows what the modern compute can achieve
Depth-aware light injection in TypeGPU
I got a 448x448 monocular depth model down to ~8 ms on my M4 Pro across ~250 dispatches, which is fast enough to use in realtime :D
Since the inference is written directly in TypeGPU, I can just feed the depth buffer straight into the lighting pass. It never has to leave the GPU or go through any extra synchronization/interop step
Inference, lighting and draw all go through the same command encoder.
Foveated streaming to my AVP in Bengaluru, India from a PC sitting in Southern California and coordinating with the product manager in Germany. Past couple months have been fun!
They finally added Snapping to RCP 🙌
So much quality of life improvements. Loving all the updates so far.
Wonder why the keynote was so subtle. So many significant improvements and additions in visionOS 27 that never got a mention. Reverb Mesh API has me very excited along with Gaussian Splat render.
Let’s do this #wwdc
1 day to go. 🍎🚀
Fireside chat at the #WWDC26 Watch Party tomorrow:
@_kjagadish & @raksha_rao - the indie duo behind Katha Room, a 2026 Apple Design Award Finalist.
Indian storytelling. Apple's global design stage. Their story.
See you tomorrow at Swiggy HQ. 👀
My one wish with visionOS 27 is support for Kiosk Mode. Having given more than 10,000 demos with about 10 headsets, I know the limitations. Wrote a lot of custom code, paired iPad experiences etc but nothing beats a native Apple support. Give me ability to play/pause a video or an experience from a paired iPad.
My one wish for visionOS 27 is better Kiosk mode. I haven’t spoken much on social media about this but we have given more than 10,000 demos with about 10 devices in the past year. Have written a lot of custom code, iPad apps etc to make it easy to support and quick but I know the limitations. Also, nothing beats Apple supporting it natively. Give me ability to play, pause a video or an experience from a paired iPad. One device consumed by many is the only way Vision Pro will work in the immediate future and Apple needs to support this better. Looking forward.
We pushed an update to Immersive India last week. Moved from showing Cityscapes to short films. Grateful to Apple for the feature again. Please watch it if you haven’t already. These films were produced with a lot of heart and shows India in a way never seen before.
Foveated Streaming from Unreal to Apple Vision Pro using OpenXR including bi-directional messaging to change colors. Took me couple days but opens so many possibilities now!
A milestone indeed !
Not long ago, we thought this might be impossible. But here we are: our Gaussian Splatting videos are now streamable, just like regular video.
No download, no app. 4DGS plays instantly in the browser on headsets, phones and laptops, with no limits on splats or length. Give it a try: store.gracia.ai/
This is my go to page to learn visionOS. Definitely recommend supporting them 🙌
If you appreciate the work we do at Step Into Vision, please consider supporting us on Ko-fi. We use this money to pay for infrastructure and fund our community work. Any amount helps!
ko-fi.com/stepinto
Tried this and within couple minutes had some seriously impressive splats to view. Glimpse of the future.
Links to the published paper and GitHub published code are here if you want to try at home:
apple.github.io/ml-sharp/