From prosthetics to Harvard: the arc of a robotics researcher
Matias's path into robotics started at Imperial College with a simple fascination: the interface between machines and humans. Prosthetics led to exoskeletons, exoskeletons led to ETH Zurich (arguably the world's top three university for robotics), and a project using Meta glasses to decode human emotions broadened his view of how machines can read and respond to people.
At Harvard, those threads come together. His PhD focuses on exoskeletons that help people with Parkinson's disease walk, combined with foundation models that let the device learn from whoever is wearing it. The goal is a device that adapts seamlessly, whether the user is a healthy 20 year old or a 70 year old managing a progressive neurological condition.
If I give a device to someone who is 20 and healthy, or to someone who is 70 with Parkinson's, the device should just learn how that person walks and adapt. That's what we're building toward.
Matias, Harvard PhD Researcher
Humanoids vs exoskeletons: the hype gap
The robotics conversation in 2025 is dominated by humanoids. Tesla's Optimus, Figure, Boston Dynamics. The sci-fi appeal is obvious and so is the venture capital. But Matias makes a case that the biggest near term impact will come from somewhere far less cinematic.
1 in 3 Europeans will be over 65 by 2050
47% of internet traffic is already AI agents and bots
86% of humanity has never used AI
The demographic case for exoskeletons is stark. With Europe's population aging rapidly, the burden on healthcare systems is growing in ways that are already straining budgets. A wearable device that helps an elderly person move independently, walk safely, or recover from injury is not a sci-fi product. It's a healthcare necessity arriving on a very clear timeline.
The humanoid question is more complicated. Matias is thoughtful here rather than dismissive. The appeal of a general purpose robot that can do many things is real and that's exactly why investment is flowing. But the counterargument is just as valid: if you need one very specific task done reliably, a purpose built robot with wheels is probably better than a humanoid that needs to learn everything from scratch.
Humanoids
- General purpose, many tasks
- Enormous investment and hype
- Longer path to deployment
- Best suited to unstructured environments
Purpose built robotics
- Specialized, highly reliable
- Faster time to market
- Immediate clinical and industrial uses
- Lower cost, lower risk
"Robotics is going to boom," Matias says simply. "The question is which form it takes first."
The blue collar conversation nobody is having
Most of the public anxiety about AI and jobs focuses on white collar work: writers, analysts, lawyers, junior consultants. But Matias and the Signal hosts spend time on a bigger and largely ignored question: what happens when robotics reaches blue collar work at scale?
The dynamic is structurally different. In white collar roles, AI tends to make teams smaller rather than eliminating them. The org chart shifts from a triangle (lots of juniors, fewer managers, one leader) toward a diamond shape: fewer people at the bottom because each person does more with AI, and therefore fewer middle managers needed to coordinate them.
Blue collar automation does not follow the same pattern.
For white collar jobs, the current AI trend looks like a reduction in headcount. For blue collar, in a lot of cases, you could see it as a replacement.
Matias
Why blue collar robotics is the bigger story
- Roughly 40% of the global workforce is in blue collar roles
- The economic logic is direct: if a robot is cheaper, companies will choose the robot
- Past industrial revolutions moved workers from blue collar to white collar, but that transition assumed white collar jobs were plentiful and stable
- Today, white collar jobs are also under pressure, closing the traditional escape route
- Reskilling at the scale needed is something governments are not yet taking seriously enough
There is an optimistic version of this story. A lot of blue collar work involves conditions that are genuinely dangerous, physically punishing, or deeply undesirable. Automating fulfillment centers or heavy manufacturing might elevate human conditions rather than simply eliminating livelihoods. But that optimism only holds if there are meaningful places for those workers to go next, and right now there is no clear answer to that question.
"These are important discussions that governments have to have," Matias says. The hosts agree, adding that universal basic income and serious reskilling infrastructure are conversations that cannot wait for the technology to arrive before starting.
Why Europe's best engineers keep leaving
Matias is European. He did his undergraduate at Imperial, his research year at ETH Zurich, and then went to Harvard. It's a journey that maps a pattern many European engineers follow, and he's honest about why.
It's not the quality of the education. European universities, especially in engineering, are excellent and often less expensive than their American counterparts. The graduate cohort Matias encountered in Europe was, by his account, highly demanding and deeply talented. But something shifts after graduation.
A lot of people who graduate in engineering look at the job market and see I can earn maybe 30 or 40 thousand as an engineer, or almost double in finance and consulting. For a lot of people, that's a very simple decision.
Matias
The talent doesn't disappear. It gets absorbed into consulting and finance, where many of these engineers stay permanently. The skills transfer. The passion for engineering does not always survive the economics.
In the US and Switzerland, engineering is compensated at a premium because the companies building on that engineering reach global markets and generate the revenue to pay accordingly. Europe has not yet built enough of those companies at scale. The result is a self reinforcing cycle: less investment in deep tech companies means lower salaries for engineers, which means the best engineers go elsewhere, which means fewer world class companies get built.
The fix is not simple. It requires more ambitious founders willing to build large companies in Europe, more capital willing to back them over longer horizons, and a cultural shift toward valuing engineering as a career path on par with finance. As Andrea puts it: "Europe would need to rethink this. If our best talent like Matias is having his best years of research done in the US, he'll end up hopefully creating a great company, but it's going to be based in the US."
Key takeaways from this episode
- Exoskeletons and purpose built robotics will likely reshape more lives in the near term than humanoids, especially as Europe ages
- Foundation models for robotics are the next frontier: devices that learn from individual users rather than being pre-programmed for average cases
- Blue collar automation follows a different logic from white collar AI. It can be replacement rather than reduction, and governments need to prepare now
- The white collar transition path that historically absorbed displaced blue collar workers is itself under pressure, making this moment genuinely different from previous industrial shifts
- Europe produces excellent engineering talent but loses much of it to better compensated markets. Fixing that requires more ambitious company building at home
- Data privacy in robotics and wearables is a real and unresolved issue, especially for devices that learn from personal movement and health data
- The AI hype cycle has a blind spot: the robotics advances happening in labs right now will drive some of the largest GDP impacts of the coming decade
The next Matias is out there.
Are you on the list?
Nova is a merit-based network of over 25,000 top professionals across 80+ countries. We find and recognize exceptional people before the world does. If you're building something remarkable or want to be around those who are, this is where to start.
