Bingyi Kang

@bingyikang

growing alongside @amilabs, prev: created and led the Depth Anything series

USA

bingyikang.com

Replies Reposts Media

Pinned
Bingyi Kang
@bingyikang
Mar 20
Excited to finally share that I’m part of the founding team at @amilabs. After many detours along the way, it’s always been about intelligence. Grateful to be doing this with @sainingxie, @ylecun, and many old and new friends.
AMI Labs
@amilabs
Mar 10
Advanced Machine Intelligence (AMI) is building a new breed of AI systems that understand the world, have persistent memory, can reason and plan, and are controllable and safe. We’ve raised a $1.03B (~€890M) round from global investors who believe in our vision of universally
26
Bingyi Kang
@bingyikang
Aug 31
Welcome to apply and drop me a note if interested!
Jihan Yang
@jihanyang13
Aug 31
New positions open! We’re looking for full-time researchers and interns with backgrounds in geometry and 3D vision to join our team at AMI Labs. The roles can be based in NYC, Paris, Montreal, or Singapore. If you’re interested, apply through the link below:
3
Bingyi Kang
@bingyikang
May 27
3D has always felt essential for visual intelligence that truly works in the real world. The challenge is making it useful in a simple and scalable way. Cambrian-P shows a surprisingly strong signal: adding pose grounding to video understanding substantially improves spatial
Jihan Yang
@jihanyang13
May 26
Camera pose matters for video understanding! Today's MLLMs excel at recognizing activities, but still struggle with the underlying space and ego/object dynamics in video. We trace this gap to a missing piece: camera pose. Introducing Cambrian-P: a multimodal LLM natively
Bingyi Kang
@bingyikang
Dec 11, 2025
Need to run DA3 on super-long videos or small GPUs? We are releasing DA3-Streaming (code) — a memory-efficient inference pipeline via chunk streaming. Built on VGGT-Long with: • Triton-optimized chunk alignment • Streaming-ready workflow 👉 Process a 20-min video in < 1 hour
4
Bingyi Kang
@bingyikang
Dec 9, 2025
We are releasing the Visual Geometry Benchmark (code + data) behind Depth Anything 3! Multi-view depth benchmarks often overweight overlapping regions due to higher point density. Ours treats the whole scene equally, providing a rigorous test for true geometry quality and
Depth-Anything-3/docs/BENCHMARK.md at main · ByteDance-Seed/Depth-Anything-3
From github.com
5