PhD · MBA · Lecturer
I am Alexander Hermann. I connect rigorous computational science and AI with human-centered innovation, and help teams turn complex problems into results they can trust.
My work sits at the intersection of computational science and innovation. On one side, I develop simulation and AI methods that describe how the physical world behaves, from the failure and degradation of materials to biodegradable medical implants. On the other, I use design thinking and innovation management to turn that knowledge into decisions, products, and ventures.
The common thread is trust. Whether it is a simulation, an AI agent, or a business decision, the goal is a result that holds up under scrutiny. My recent work on agentic AI for engineering, the OASiS framework, is built on exactly this idea: let an AI agent do the heavy lifting, but verify every result before anyone relies on it.
I completed my doctorate in computational mechanics with distinction (summa cum laude) at Hamburg University of Technology, where I now teach, lead research, and am working toward my Habilitation. Alongside my engineering background, I hold an MBA in Technology Management, and my current Doctorate in Business Administration at Liverpool John Moores University studies exactly this question of trust: how it shapes organisational resilience under pressure, and how a practical decision-support tool can help teams respond and learn after critical incidents. As a University Innovation Fellow of the Stanford d.school, I enjoy bringing that mindset to teams that want to innovate with confidence.
Where I have worked and studied, at a glance.
My research develops methods to model how materials and systems behave, and to make advanced simulation faster, more reliable, and easier to use.
I build nonlocal models that capture how materials deform, crack, and fail, including efficient methods for problems that reach beyond a finite boundary. This work makes fracture and wave simulations both more accurate and more affordable.
I model magnesium-based implants and biomaterials for medicine, from bone screws to cardiovascular stents, so devices can be safer, patient-specific, and able to dissolve safely once they have done their job.
I bring agentic AI and machine learning to scientific computing, from forecasting material degradation to letting AI agents run and verify full simulations. The focus is always on reliability, so results can be trusted in practice.
Describe a physics problem in plain language, and OASiS chooses the right solver, runs the simulation, and checks the result before it comes back to you. It connects nine open-source simulation codes and works inside AI assistants like Claude.
Two areas where my models turn difficult physical problems into results you can see and build on.
Some implants can do their job and then dissolve safely in the body, so a patient is spared a second operation to remove them. I model how magnesium-based implants, from bone screws to cardiovascular stents, lose material and strength over time. This helps make these devices safer and tailored to each patient.
More in my publications and open-source code on GitHub.
The brain depends on a close interplay between nerve cells, their supporting cells, and a fine network of blood vessels. I build multiscale models that show how blood flows through these vessels and how oxygen reaches the surrounding tissue, which helps researchers understand both healthy brain function and what changes in disease.
Published in the Journal of Mathematical Biology; code on GitHub.
Agentic AI that operates nine simulation codes and verifies every result.
A memory-efficient key-value cache for large language models.
Multiscale brain model coupling nerve cells, blood flow, and tissue transport.
3D model of how biodegradable magnesium implants degrade over time.
Parameter fitting and calibration for peridynamic material models.
A fast voxelizer that turns 3D geometries into simulation-ready grids.
Voxelizers, physics-informed networks, peridynamic solvers, and more.
Strong research only creates value when it reaches the real world. That is why I invest as much in the how of innovation as in the science itself.
As a University Innovation Fellow of the Stanford d.school, I have led design thinking workshops that help teams start from real user needs and move quickly to tested prototypes. Through the AI Startup Hub Hamburg ideation program (Winter Batch 2024) I worked on an AI-driven biosimulation concept and gained hands-on experience with early-stage validation and pre-seed fundraising. My MBA in Technology Management and my ongoing DBA on trust in decision making round out this side of my work.
Whether you want to apply AI and simulation to a hard problem, run an innovation workshop with your team, or get an expert opinion, here is what you can bring me in for.
I bring agentic AI, machine learning, and high-performance simulation to real engineering and materials problems, from a first prototype to a workflow your team can rely on.
Hands-on workshops and training based on the Stanford d.school method. I help your team turn user needs into tested ideas, and coach you to run the process yourselves.
A technical second opinion on your modeling and simulation, ongoing advisory, and talks for teams, events, and universities.
Tell me a little about it. I read every message myself and usually reply within a few working days.
Whether it is a research collaboration, a consulting question, a workshop, or a talk, I would be glad to hear from you.
