Vickram Rajendran

making machines learn @ Applied Intuition

[Malhar Patel @malharhar]

I wanted my first public video to be learnings from the past 7.5 years here at @AppliedInt. Hopefully it’s useful to the next set of builders 🐻

[Vickram Rajendran @Vickitidrum]

@changanvr Been super excited to see what you guys have been up to, would love to chat more with you guys on Thursday!

[Vickram Rajendran @Vickitidrum]

@ArunabhMishra8 See ya there!

[Chuhan Zhang @ChuhanZhang5]

A SINGLE encoder + decoder for all the 4D tasks!
We release 🎯 D4RT (Dynamic 4D Reconstruction and Tracking).

📍 A simple, unified interface for 3D tracking, depth, and pose
🌟 SOTA results on 4D reconstruction & tracking
🚀 Up to 100x faster pose estimation than prior works

[Vickram Rajendran @Vickitidrum]

Kind and wicked learning environments was such a key insight for me

[Vickram Rajendran @Vickitidrum]

Aka how to get good at anything

Nice post once again that I resonate with a lot

[Eryk @Ambisinister_]

How to win at games https://t.co/D8Bmu7Uagb

[Applied Intuition @AppliedInt]

We’re trading simulation labs 🤖for surfboards 🌊.

Meet Chief Scientist @Wei_ZHAN_ +@Vickitidrum - some of the brains behind our Self-Driving System (SDS) at #ICCV2025 🌺.

Find us at Booth #127 — talk autonomy, ask questions, or just say hi. 👋.

👀 ➡️https://t.co/waAcggdV7h

[Vickram Rajendran @Vickitidrum]

@vikhyatk Similar to the first chain of thought papers https://t.co/DOcSb1iV1G

[Vickram Rajendran @Vickitidrum]

moondream may be the best model per gpu hour in the world

[Vickram Rajendran @Vickitidrum]

Autonomy for everyone

[Applied Intuition @AppliedInt]

Today, we’re officially announcing our Self-Driving System for Automotive — our L2++ ADAS product for passenger vehicles. 🚀

Meet the system under the hood: https://t.co/PG8AxbOwGF
#autonomousvehicles #vehicleintelligence #adas

[Applied Intuition @AppliedInt]

Today, we’re officially announcing our Self-Driving System for Automotive — our L2++ ADAS product for passenger vehicles. 🚀

Meet the system under the hood: https://t.co/PG8AxbOwGF
#autonomousvehicles #vehicleintelligence #adas

[Malhar Patel @malharhar]

😀🚗
https://t.co/mPhwOxmADc

[Vickram Rajendran @Vickitidrum]

OpenAI + vehicles oh my

Been fun working with the OpenAI folks and lots more to come (:

[Applied Intuition @AppliedInt]

Applied Intuition is collaborating with @OpenAI to advance next-generation, AI-powered experiences in vehicles worldwide.

Learn more about this collaboration: https://t.co/7tzIi1CXfU

#AI #vehicleintelligence #OpenAI

[Vickram Rajendran @Vickitidrum]

@umbhutta super exciting, congrats! Is there a link to the paper?

[Vickram Rajendran @Vickitidrum]

@EugeneVinitsky Incredible work on the engineering, reward function design, advantage filtering!!, and just belief that this would work if you scaled hard enough! Really appreciate the detail on all the modeling and training strategies here too (:

[Vickram Rajendran @Vickitidrum]

@francoisfleuret @deepseek_ai https://t.co/fEPyU3jP3J great more detailed overview on changes they made through their papers

[Vickram Rajendran @Vickitidrum]

@brog8264 Hell yea dude nice work!!!

[Vickram Rajendran @Vickitidrum]

Best overview of modern LLMs I've read, following Deepseek going from llama-2 performance to easily open source SOTA in essentially a year. highly recommend

[Eryk @erykbanatt]

Understanding Modern LLMs via DeepSeek - Part 2

https://t.co/VfFRxNmIOW

As before, an annotated reader for the DeepSeek papers. Please let me know what you think!

[Vickram Rajendran @Vickitidrum]

@songyoupeng Is there any training code on the github? Looks like just inference right now!

[Vickram Rajendran @Vickitidrum]

@songyoupeng Very cool work! Any thoughts on using focal lengths/metric depth/flat plane, or metric depth from another model instead of lidar for the prompting here? How accurate does the prompt need to be, is lidar(even cheap) the only way?

[Vickram Rajendran @Vickitidrum]

@HaotongLin @songyoupeng 🙏🙏 thank you!!! Yeah, definitely curious how other prompting will look for this!