@omnia_io @IlirAliu_ That’s another solid source, a lot of promise there.
English
Deplace AI
10 posts
Deplace AI
@DeplaceAI
The Physical AI deployment layer
Paris, France Katılım Nisan 2025
23 Takip Edilen99 Takipçiler
@IlirAliu_ @DeplaceAI Pair it with data coming from AR sessions. Imagine how much you can do with POVs from normal people wearing smart glasses
English
You will soon be able to teach robots what human are doing… using natural language.
I spoke to one of the founders a couple of months ago on my podcast:
An API that takes raw human videos and returns detailed language annotations describing the motion.
@DeplaceAI is building something wild:
Not just basic labels, it captures:
✅ motion semantics
✅ relative positions
✅ cause and effect
✅ task outcomes
✅ and more
It’s built on top of research in point tracking, segmentation, and learning from demos;
all to make it easier to train robots and embodied agents without manual labeling.
They’re offering early access and sharing some of the datasets they collected via their global network of video collectors.
Thanks for sharing, @MilcentPedro !
🔗 forms.gle/sSHTPBDVkBtnHh…
One of the most interesting motion-to-language interfaces I’ve seen.
If you’re working in robotics, vision, or LfD, it’s worth checking out.
English
Thank you for sharing our work @IlirAliu_
Human motion holds a lot of knowledge to fuel Physical AI!
Ilir Aliu@IlirAliu_
You will soon be able to teach robots what human are doing… using natural language. I spoke to one of the founders a couple of months ago on my podcast: An API that takes raw human videos and returns detailed language annotations describing the motion. @DeplaceAI is building something wild: Not just basic labels, it captures: ✅ motion semantics ✅ relative positions ✅ cause and effect ✅ task outcomes ✅ and more It’s built on top of research in point tracking, segmentation, and learning from demos; all to make it easier to train robots and embodied agents without manual labeling. They’re offering early access and sharing some of the datasets they collected via their global network of video collectors. Thanks for sharing, @MilcentPedro ! 🔗 forms.gle/sSHTPBDVkBtnHh… One of the most interesting motion-to-language interfaces I’ve seen. If you’re working in robotics, vision, or LfD, it’s worth checking out.
English
Deplace AI retweetledi
First day at @VivaTech & @NVIDIAGTC with an interesting keynote from Jensen and with #robotics & Physical AI companies represented.
Not every day you see @Tesla_Optimus and @engineered_arts under the same roof!
English
Deplace AI retweetledi
We have been seeing great progress in using human demos in Physical AI models.
This paper by @ryan_hoque, @peide_huang, and team shows how a large-scale dataset can help scale this: 800+ hours of human demos for manipulation.
📄: arxiv.org/abs/2505.11709
English
Deplace AI retweetledi
Great new paper by @LihanZha, @ApurvaBadithela, @mzhangio, @justinlidard, @Majumdar_Ani, @allenzren, @shahdhruv_, and team.
As we scale data collection for robotics, it must be done intelligently by choosing which data variations to focus resources on. This paper provides a great foundation for that, will definitely be using it at @DeplaceAI 📈
Check it out: factored-data-scaling.github.io
English
Deplace AI retweetledi
DexWild is a great new paper by @pathak2206, @mohansrirama, @_tonytao_ , and team. It introduces the DexWild system to collect human demos at scale, enabling robust #robot policies that generalize to novel environments, tasks, and embodiments. Highly relevant to our data work at @DeplaceAI 🤖
English
Deplace AI retweetledi
New paper from @pabbeel, @junyi42 & team! They show how visual imitation enables context-learning for humanoids, using a Real2Sim2Real pipeline that turns videos into transferable skills 📹
👉 videomimic.net
English
Deplace is glad to be a part of the NVIDIA Inception Program, accelerating the future of Physical AI with powerful tools and support from @nvidia!
Pedro Milcent@MilcentPedro
We’re happy to join the Nvidia Inception Program! Excited to accelerate the development of our robotics #data solutions with the support of @nvidia. They have been doing great work supporting Physical AI innovation🤖🚀
English
Deplace AI retweetledi
🎥 Kicking off our first 1-minute Physical AI paper review with: Physics-Driven Data Generation for Contact-Rich Manipulation
A great paper by @lujieyang98, @hjterrysuh, @tarikkelestemur, @RussTedrake, Tong Zhao, Bernhard Paus Græsdal, Jiuguang Wang & Tao Pang.
📄 Check it out: arxiv.org/pdf/2502.20382
English
Deplace AI retweetledi
Chatted with @IlirAliu_ about what we’re building @DeplaceAI, how our team got together, and the founder life in the Paris #robotics scene. Check it out 👇
Ilir Aliu@IlirAliu_
🎙️ Episode #62 of @builddeeptech! With @MilcentPedro, Co-Founder of @DeplaceAI; data collection & curation for robotics and physical AI: In this episode, I chat with Pedro Milcent, a French-Brazilian founder building something new in the world of robotics and AI. Pedro’s story starts unconventionally, as he finished high school at 14, went straight to law school, then pivoted to business, earning a full-ride scholarship to USC before diving into the startup world. Pedro’s first startup, Payby, tackled payments in Brazilian restaurants. Now, he’s co-founded Deplace AI, a company solving one of the hardest problems in robotics: large-scale, low-cost data collection for physical AI. Think human demonstrations, imitation grippers, motion capture, and making it all accessible. We talk about building a team with deep expertise, landing their first client just months in, and what Pedro sees coming next for physical AI in Europe and beyond. He’s got bold views and clear insights, hope you enjoy the episode as much as I did recording it. Episode 62 – Build Deep Tech "Physical AI Has a Data Problem" With Pedro Milcent, Co-Founder of Deplace AI 🎥 YouTube: youtu.be/VdMENTpAH4s 🎧 Spotify: open.spotify.com/episode/1bo44t… 🎧 Apple: podcasts.apple.com/us/podcast/bui…
English