
Staff Research Scientist @3rddimensionai Previously: @Niantic, intern @Adobe Computer Vision, Graphics and ML. Opinions are my own.
London, England
Joined March 2012
- Tweets89
- Following2K
- Followers186
- Likes443
Áron Monszpart retweeted
Two years ago, @pyetras and I started @3rddimensionai with a simple conviction: the future of simulation would be grounded in reality. Only then could autonomous systems be validated in sim, at scale.
Today, I’m excited to share that @3rddimensionai is being acquired by @Rivian
The real world is incredibly complex, with a long tail of rare events that makes validation nearly impossible at scale. Instead of trying to recreate that complexity by hand, we believed we could start with reality itself — reconstruct the world from sensor data, then make it controllable, editable and generative.
Over the last two years, we built that technology and assembled one of the most incredible teams working in spatial AI. Rivian is a natural fit for our next chapter.
It was a competitive process, but we ultimately chose Rivian for four reasons:
- The product is simply incredible.
- We’ll be 100% focused on our original thesis: how do we use simulation to validate autonomy at scale? We’ll get to put what we’ve built directly in the loop with real vehicles, real-world data and an autonomy stack.
- The Rivian team is incredibly ambitious. This already shows in the quality of the vehicles, and the autonomy available to date.
- James has built a no nonsense, first principles team focused on the task at hand.
On a more personal level, I’ve long admired @RJScaringe grit and determination as a founder, and I’m looking forward to learning from him.
There are so many people who got us here.
@pyetras, my co-founder, has been an incredible partner from day one, and never stopped pushing our vision forward and the bar higher technically.
Christian, our Chief Researcher, elevated how we think about research and what we should be ambitious enough to tackle.
And the entire @3rddimensionai team: many of you left bigger, more established, better-funded companies to take a chance on this. I can’t thank you enough. I’m incredibly proud of this team, and very happy that we get to stay together and keep doing what we love.
As a first-time founder, I’m incredibly grateful for the backing of our investors. Their support gave me the confidence to take risks, set ambitious goals, and be honest when something wasn’t working.
@asenkut @DetweilerJames at @felicis gave us the backing and legitimacy to think much bigger. Fouad at @Soma_Capital wrote me my first check as a founder. The teams at @MVP_Ventures and @AbstractVC rode out the highs and lows and rolled up their sleeves when we needed them. Thank you.
And to my wife, @clairekart , and all the partners and families behind our team: startups have a way of becoming part of your lives whether you signed up for them or not. Thank you for the late nights, the uncertainty, and all the times the company came home with us. I’m confident this next chapter is a positive one for our employees and their families.
Finally, thank you to @RJScaringe, James, Ben and everyone at @Rivian for believing in the team and what we’ve built.
We have a lot to do. I’m excited to get started.
techcrunch.com/2026/10/22/ri…
Áron Monszpart retweeted
"World models" is one of the hottest terms in AI right now, but one of the least well-defined.
Our new blog digs into what we mean by it, the challenges in building one, and why we believe it's the key to solving physical AI's biggest bottleneck.
Read: thirddimension.ai/blog/posts…
Áron Monszpart retweeted
Huge thanks to Christian Rupprecht (@Oxford_VGG & @3rddimensionai) for an incredible first keynote today!
"3D Computer Vision: Done and Dust3d?"
Áron Monszpart retweeted
Reminder: Our CRO Dr. Christian Rupprecht is delivering the keynote at @3dvconf, exploring whether 3D as we know it is about to become obsolete.
And come find our team after - CEO & Co-Founder @tolgakart and Founding Engineer @EsDaSuKa would love to hear your thoughts 👋
Áron Monszpart retweeted
We'll be at @3dvconf later this week. Here's where you can find our team.
Join us Saturday evening for Render & Recharge, Third Dimension's gathering for the 3D community.
👉 Space is limited, grab your spot early: luma.com/2drnsvbj
Áron Monszpart retweeted
“Something happens in the real world… I’m back at my desk in a full 3D environment in under two hours. Now I can build confidence that what I’m trying will translate into behavior in the real world.”
Neural simulation is starting to close the robotics domain gap. Conversation with @3rddimensionai's CEO
@tolgakart on neural sim + Gaussian splatting 👇
Áron Monszpart retweeted
Big news: Dr. Christian Rupprecht (@Oxford_VGG, @CVPR 2025 Best Paper) is joining Third Dimension as Chief Researcher!
Christian will lead our efforts to advance our world modelling and spatial generation technology into unparalleled simulations and worlds.
Here’s to turning research into reality 🔥
Áron Monszpart retweeted
Robotics simulation is broken. Today, Third Dimension launches SuperSim to change this.
Using state of the art 3D reconstruction and world modeling, SuperSim reconstructs reality with sensor data and generates infinite, hyper-realistic worlds.
Learn more: bit.ly/3Kl4TnJ
Áron Monszpart retweeted
We're heading to #ICCV25 next week! 🌺
Here's where you can find the Third Dimension team 👇
Áron Monszpart retweeted
Happy to share the work from my internship at @nianticspatial: ACE-G, which we will present at ICCV next week in Hawaii 🌴
ACE-G introduces large-scale pre-training to previously mainly per-scene-optimized scene coordinate regressors. (1/n)
Áron Monszpart retweeted
I was at @foxglove Actuate this week and it really felt like a who’s who of physical AI. I truly enjoyed meeting some of the people I had been looking up to, and definitely made new friends. Congratulations to @adrianmacneil and team for bringing together such an impressive group of people.
Some takeaways inspired by this week’s conversations:
Robots are having a renaissance
Compared to even a year ago, I see more new companies working on different aspects of robotics, from food delivery to factory operations, to construction. And then there are robots that want to do it all.
Neural Simulation vs. Rendering Engine Simulation
As a founder whose company is developing a neural simulator I’m admittedly biased, but I have high conviction that combining diffusion and reconstruction is the right approach for simulation. A key advantage: Our use of Gaussian Splatting solves the domain gap problem - ie. Third Dimension’s simulator looks like you're stepping into an actual video. Not only does your simulation look like a video, but all other actors in the scene behave exactly like they did in real life - this is just not possible with classical rendering engines.
On the other hand, video game-based simulators are good at providing ground truth, such as the automatic labelling of datasets, which is important for autonomy and robotics training. The downside lies in how expensive they are to build and that they just don’t look real.
Both…For Now
The realization for me is that there is room for both, for now. You want to get pixel perfect ground truth annotations? Tap into your classical sim engine. Want to make better use of your ever growing real world data? Neural simulation is your friend.
Áron Monszpart retweeted
Splats are great, but wouldn’t it be wonderful if we could transform them into the stuff of dreams? Introducing our #CVPR2025 paper ✨Morpheus✨!
Morpheus lets you change the shape and appearance of 3D Gaussian splats with a few words!
nianticlabs.github.io/morphe…
(1/6)
Áron Monszpart retweeted
Niantic is hiring PhD students for internships in 2025:
job-boards.greenhouse.io/nia…
Áron Monszpart retweeted
We wrote a text on how we envision the future of visual mapping: nianticlabs.com/news/largege…
So far, we have trained ~50 million ACE networks. 😎 Together, that's 150 trillion parameters... Not bad! But the future lies in going from local small models to a global large model.
Áron Monszpart retweeted
We updated the #ACEZero repository to make it compatible with newer versions of @nerfstudioteam. We added some functions to make it easy to train Gaussian splats on top of ACE0.
See our FAQ: "I want Gaussian splats from my images. What do I need to do?"
github.com/nianticlabs/aceze…
Áron Monszpart retweeted
HUGE shoutout to our awesome #ECCV2024 🇮🇹 Program Chairs for all their efforts!
@eliricci_ @gulvarol @orussakovsky @stefanroth @SattlerTorsten
Áron Monszpart retweeted
We'll be at poster 105 this afternoon! In the meantime, you can try an interactive DEMO of a lightweight mobile version of HandDGP directly on ANY device with a browser! nianticresearch.8thwall.app/… #ECCV2024
👋Would you like more accurate 3D camera-space hand mesh predictions? Check our new #ECCV2024 work: "HandDGP: Camera-Space Hand Mesh Prediction with Differentiable Global Positioning". Project page: nianticlabs.github.io/handdg… (1/5)
Áron Monszpart retweeted
We have released the code and model weights on github.com/nianticlabs/HandD… #ECCV2024
👋Would you like more accurate 3D camera-space hand mesh predictions? Check our new #ECCV2024 work: "HandDGP: Camera-Space Hand Mesh Prediction with Differentiable Global Positioning". Project page: nianticlabs.github.io/handdg… (1/5)
Áron Monszpart retweeted
Attending @eccvconf and looking for PhD or postdoc positions in 3D reconstruction or visual localization? Happy to talk to you :)

















