About us
As AI systems grow in scale and complexity, their infrastructure must evolve. We are building the optical technology that will enable data to move faster and more efficiently than ever before.
Dipole Labs was founded in 2026 and is backed by Y Combinator, with headquarters in Boston and R&D centers in San Francisco and Lausanne, Switzerland. The compute bottleneck is here now, and we are building the way past it. If you want to build at that frontier, let's talk.
PhD in Physics and Mechanical Engineering from ETH Zürich. Former postdoctoral researcher at Harvard, where he explored new ways to use light for computing. At Dipole, he is applying that expertise to rethink how AI systems connect and scale. Has the habit of getting to the bottom of things: steady, Swiss rigor, and optimism that doesn't run out.
PhD in Applied Physics and Materials Science and Engineering, EPFL. Former postdoctoral researcher at Harvard, where he worked on novel optical probes for quantum materials. Has a habit of finding the craziest solutions, with explosive creativity paired with a sharp sense of humor.
PhD from EPFL, where he demonstrated the first room-temperature excitonic transistor. Dimitri led a large R&D division, taking optics and photovoltaics products from proof of concept to industrialization. Has a habit of turning physics into products, with calm structure and endurance-athlete stamina.
We are hiring engineers and scientists across process, photonics, optics, electronics and systems.
See open roles →
AI is arguably the most consequential thing humanity has ever built. It could solve problems that have always felt out of reach, from curing diseases to feeding the world. That future, as crazy as it sounds, is possible in our generation. But that future has to run on physical infrastructure, and today the infrastructure can't keep up with the progress. The infrastructure IS the bottleneck.
Today, that infrastructure runs on silicon. Light already carries the world's information, because it is the fastest medium we have and it barely loses energy over distance. But the moment that light reaches a silicon chip, we convert it into electrons. Every conversion burns power, adds latency, and throws off heat. A single AI prompt triggers thousands of these conversions, so we pay that tax billions of times a day.
We already trust light to cross oceans. The question is why we still give it up the moment it matters most.
Light already moves in and out of every data center. We are building the light-based network that preserves it in between and enables the AI revolution.
See open roles