RESEARCH SPOTLIGHT
Hi, I am
Peter Samaha
HOST INSTITUTE
Hewlett Packard Enterprise
PROJECT TITLE
Algorithm-Hardware Co-Design of Non-von Neumann Architectures for Artificial Intelligence
KEY WORDS
Algorithm-Hardware Co-design, Non Von Neumann Architectures, Unconventional Computing, Photonics, Energy-Based Computing
RESEARCH AIM
My research aims to co-design analog systems and unconventional computing models so that we can unlock fundamentally new paradigms of machine intelligence beyond traditional von Neumann limits.
RESEARCH OVERVIEW
Current artificial intelligence is constrained by the rigid digital abstraction of standard computing, where shuttling data between separate memory and processing units creates an unsustainable energy bottleneck. My research bypasses this by completely shifting the computational paradigm. Rather than writing heavy algorithms and forcing them through inefficient digital hardware, I co-design the algorithms and the hardware simultaneously.
I use analog photonics and mixed-signal electronics to build systems where the hardware natively performs the math. Ultimately, this approach turns the laws of physics themselves into the computational mechanism, unlocking a vastly more efficient, physics-driven paradigm of machine intelligence.
RESEARCH CHALLENGE
The primary challenge my research aims to solve is the fundamental energy wall and data-movement bottleneck—often called the von Neumann bottleneck—that currently limits the scaling of artificial intelligence. Today's standard AI models running on conventional hardware consume massive amounts of power simply shuttling data back and forth between memory and processors. Furthermore, by confining machine learning strictly to digital abstractions, we ignore the immense computational potential of physics itself. My work addresses this by completely shifting the paradigm: rather than trying to force heavier algorithms through inefficient digital architectures, I co-design analog substrates and unconventional, energy-based models. This allows the natural physical dynamics of the hardware to compute natively, entirely bypassing traditional digital bottlenecks and unlocking vastly more efficient, physical embodiments of intelligence
RESEARCH INNOVATION
What makes my approach unique is the bidirectional nature of the hardware-algorithm co-design, deliberately abandoning the standard digital abstraction layer. Traditionally, there is a disconnect: computer scientists design heavy AI algorithms and expect hardware to execute them, while physicists design novel optical chips and then search for algorithms that fit them. My approach bridges this gap simultaneously.
RESEARCH IMPACT
What excites me most is working at the intersection of hardware implementation - such as photonics and analog electronics - and artificial intelligence. This opens up opportunities not only to design more efficient, co-optimized systems, but also to explore different physical manifestations of intelligence across novel architectures and substrates (e.g., Ising machines, physical reservoir computing, and energy-based computing)
RESEARCH SUMMARY
My research focuses on the algorithm-hardware co-design of non-von Neumann architectures, leveraging photonics and mixed-signal electronics to pioneer new paradigms in physical computation. Traditional digital computing and the standard artificial intelligence algorithms developed for it face fundamental energy and data-movement bottlenecks. To overcome these limits, my goal is to build substrates that compute naturally alongside the specific algorithms best suited for them—rather than forcing conventional models onto new accelerators. By exploring this algorithmic-physical interface, I create unconventional platforms where the continuous dynamics of the hardware natively solve complex problems. By mapping mathematical optimization and learning tasks directly onto these systems through energy-based principles—where the physical state naturally evolves toward a solution—my work bridges hardware physics and AI to unlock fundamentally more efficient machine intelligence.
RESEARCH VISUAL
RESEARCH JOURNEY
RESEARCH SECONDMENT
Ghent University, Belgium (3 months)
During my secondment at Ghent University, Belgium, I will focus on the experimental validation of the proposed photonic systems through a series of research visits taking place in September 2026, September 2027, and January 2028. This secondment will provide valuable hands-on laboratory experience, allowing me to strengthen my experimental skills while validating the performance of the systems developed throughout my research.
LEARNING GOALS
I am particularly looking forward to gaining hands-on laboratory experience while broadening my knowledge through exposure to new research topics and ideas. By collaborating with researchers from different subfields, I hope to gain fresh perspectives that will strengthen both my research and professional development.
RESEARCH BENEFITS
These secondments will enable me to translate my theoretical research into practice by physically building, testing, and characterising the photonic circuits and systems I have designed.
CAREER DEVELOPMENT
These secondments will support my professional development by enabling me to experimentally validate my research ideas while expanding my technical expertise through hands-on laboratory experience.
MORE THAN PHD
Why did you decide to join POSTDIGITAL+ ?
I joined POSTDIGITAL+ because it offers a unique interdisciplinary research environment and the ideal combination of expertise, collaboration, and international mobility needed to support this highly cross-disciplinary research project.
What do you think is the biggest advantage of being part of the International Doctoral Network?
The biggest advantage of being part of an international doctoral network is the structural cross-pollination of ideas, methodologies, and resources that prevents you from getting trapped in a single lab's perspective. Ultimately, it accelerates your technical breakthroughs by blending international expertise, while instantly building a collaborative professional network that transitions you from a siloed investigator into a globally connected researcher.
What skill have you developed the most so far?
One of the key skills I have developed during my PhD is scientific communication and outreach, enabling me to share complex research with a wide range of audiences through effective dissemination and public engagement activities.
What are your career ambitions after completing your doctorate?
After completing my doctorate, I hope to establish a deep-tech start-up that translates cutting-edge research into innovative technologies with real-world impact.
What inspired you to pursue this research Area?
I was inspired by the idea that artificial intelligence does not have to be constrained by traditional digital computing. Instead, I am fascinated by the possibility of harnessing the natural dynamics of physical systems to perform computation, while exploring entirely new forms of machine intelligence.
What surprised you most since starting your PhD?
What surprised me most was seeing how researchers from different disciplines can work on entirely different challenges while all contributing towards the same shared goal of advancing next-generation computing technologies.