@Harwig spotted this LinkedIn post by Massachusetts-based Geisel Software about two weeks after Akida Pico was launched back in October, so we know they are definitely aware of us. They really do look like an intriguing company to keep an eye on (and to potentially partner with).
The CEO and Founder is switched on.
He was "also" at Edge A.I. Vision Alliance 2024, which is where he most probably met us (or it possibly happened earlier)
Share your videos with friends, family and the world
m.youtube.com
Only 28 subscribers, but his channel is worth going over..
Possible links to Apple? As he describes their AR headset, in one of his clips, or just general tech information?..
A focus on robotics.
The Geisel Software channel has ~5500 subscribers, still very small.
Share your videos with friends, family and the world
m.youtube.com
A space worth watching..
This is what they posted yesterday:
We’re already seeing rapid change but some of the most exciting tools are still just emerging. Here are 3 technologies we believe will become foundational in the next 5 years: 🧠 Neuromorphic computing: Chips that mimic the human brain to enable faster, more energy-efficient robotic perception...
www.linkedin.com
And other similar posts in recent weeks:
Discover the best strategies for optimizing Edge AI to enhance real-time decision-making in robotics.
buff.ly
(This February 2025 blog post by Geisel Software Head of Marketing Kristin Wattu only mentions Loihi and in-memory computing, though, with regards to neuromorphic computing).
Login to LinkedIn to keep in touch with people you know, share ideas, and build your career.
www.linkedin.com
Founder and CEO Brian Geisel founded a spin-off company last year, of which he is also CEO: Symage, Inc. by Geisel Software describes itself as “a synthetic data platform that generates real results”. It creates photo-realistic, precision-labeled synthetic image datasets to enable faster, cheaper and more efficient AI model development. After all, AI models are only as good as their training data.
Symage actually evolved from an interesting challenge in the space sector: NASA had commissioned Geisel Software “to create high-fidelity Mars synthetic data to train Mars rovers in recognizing ‘blueberries’ - small spherical nodules composed of hematite. These nodules likely formed from water beneath the Martian surface and could indicate the presence of past life”.
Training of computer vision systems used in robotic planetary explorations is very challenging given the lack of real datasets - all we have so far are 2D-images and videos of small sections of those terrains. So what Brian Geisel and his team effectively did was to create a virtual 3D-Mars environment, a simulation based on real images of the Red Planet that NASA’s Jet Propulsion Laboratory provided them with. They would then, for example, model their virtual rocks on real terrestrial rocks that looked similar to Martian rocks. Brian Geisel talks about this in the video below (of which I’ve watched the first 40 minutes or so).
It’s actually quite hilarious to hear how Geisel Software’s engagement with NASA initially started out… See also this recent LinkedIn post:
I was on Founders in Focus with Arthur Lozinski recently, and he asked, “So what’s going on at NASA? And what are these robots doing?” I had to laugh. Because the way we got involved wasn’t some polished, strategic move. It was pure chutzpah. One day, I stood up in a Geisel Software meeting...
www.linkedin.com
Brian Geisel and his team are convinced that “Synthetic Data is the Future of AI Training”, not only for space, but also here on Earth:
This is what he posted two weeks ago (sorry, my upload limit for files is exhausted…)
“Why synthetic data on Earth?” We started building it for Mars—because no training data exists there. But it turns out… Earth needs it too. When real data is too risky, synthetic steps in: ✔️ No privacy risk ✔️ No permissions ✔️ No liability Think banks. Hospitals. Logistics. Anywhere you...
www.linkedin.com
“Why synthetic data on Earth?”
We started building it for Mars—because no training data exists there.
But it turns out… Earth needs it too.
When real data is too risky, synthetic steps in:
No privacy risk
No permissions
No liability
Think banks. Hospitals. Logistics.
Anywhere you need to train AI—without compromising people.
Once you see it, you can’t unsee it.”
And this last week:
Sounds like synthetic data is definitely here to stay…
...
