You describe yourself as a creative technologist. How do you define this nomenclature?

I work at the intersection of architecture, machine learning, climate engineering, and culture. My role as a creative technologist is about using technology actively and imaginatively by breaking it open, seeing how it works, and misusing it creatively. I think of myself as someone who tinkers, almost like an engineer in a garage, except instead of hardware, I’m working with datasets and AI models. Whether visualizing a city’s future infrastructure, creating a deepfake music video, or simulating urban climate adaptation strategies, my projects combine advanced tools with systemic storytelling. The term “creative technologist” captures that interdisciplinary, tool-hacking, deeply exploratory nature of my practice.

I’m intrigued by “systemic storytelling”. Is that a new narrative in design thinking?

I hope it is, or will be soon. I live in Los Angeles, but I don’t think of narrative in a cinematic, Hollywoodian way. To me, a narrative is about systems logic. If we change one element in a system, what happens across the whole ecosystem? To me, that’s the story. In this view, architecture is about visualizing the interdependencies of our lives and expressing and solving how space interacts with this, as well as with infrastructure, ecology, and data. For example, in one of the courses I give at UCLA, I have my students trace the life cycle of a food ingredient, a tomato, for instance, from where it is grown, through its transportation, consumption, and composting. This becomes a story about consciously mapping energy, economy, and environment with a relatable emotional charge. When we begin to see our data as storytelling, architecture becomes a tool to model complexity.

I try to prepare students for the kinds of cross-disciplinary thinking they’ll need in practice. I bring in projects from my work, like simulations of tidal flow and sediment movement around islands, or energy modeling for buildings that use natural ventilation. We use game engines not for games, but to simulate climate conditions, airflow, sunlight. That way, students see how technology isn’t just about novelty—it’s a language for design. One of the biggest goals is teaching them to see architecture not as object-making, but as systems-thinking. It’s about recognizing patterns, making meaningful decisions at multiple scales, and innovating the architecture and design studios of tomorrow with that thinking.

How does adapting to technological acceleration enhance the design process?

AI doesn’t just replicate, it transforms data, evolving it to express new possibilities. When training models, the goal isn’t perfection but exploration. I’m drawn to algorithmic mutation and the way datasets can “contaminate” the design process in unexpected, creative ways. Errors or hallucinations often reveal new pathways. In my early fashion work with Esque, I trained CycleGANs¹ and later StyleGANs² on curated image datasets. The generated images were uncanny, sometimes strange, but always rich with potential. I converted them into 3D forms using tools like Clo3D and tested workflows that move from sketch to structure. That progression, from dataset to sketch to model, is foundational to my architectural workflows now. The same logic extends to urban simulations, where machine learning isn’t just a tool for optimization but a tool for imagination.

Is this imaginative way of working with technologies a form of what you call “machine agency”?

It is, yes. AI models evolve rapidly, making their integration into design both dynamic and unpredictable. This shift demands an expanded definition of the field. Contrary to what we’ve been taught about individualism, machine agency, as we move into the future, is not about who authors a design. It’s about how we guide the machine. Tools leave imprints. A paintbrush has a signature, and so does an AI model. Recognizing that signature and choosing how to wield it is part of the aesthetic. I don’t hide algorithmic traces when I work; I highlight and shape them. I even build models that reflect my preferences—my “style,” if you will. There’s a misconception that AI homogenizes design. In fact, the more you train your own models, the more distinct your results become.

You’ve also spoken about reclaiming “architectural agency”. Is training your own models in your own style part of this agency?

Yes. Architects are storytellers who envision how people will live and interact within space. They used to oversee the entire construction process. But increasingly, responsibilities are outsourced to consultants, engineers, and software. That diffusion of responsibility has eroded architectural agency. AI and automation, if approached critically, offer a way to bring some of that control back. For instance, we can use Python to parse zoning data, apply that to 3D visualization workflows in Unreal Engine, and generate speculative data-driven urban futures. Instead of relying on external consultants for analysis and modeling, we can integrate those processes into our own design practice. It’s a way of reasserting authorship in a field that’s been fractured.

I believe that design will increasingly emerge from a hybrid convergence of disciplines, tools, methodologies, and interdisciplinary collaboration where technology is the catalyst. In this, I see the democratization of information, open-source frameworks, and knowledge-sharing as critical. There’s a special opportunity now for architects to be active participants in shaping new technology.

Can you give a project example that illustrates your own architectural agency?

The Festival Trail project in Los Angeles is a great example. It was part of a city-wide initiative to reimagine urban corridors in anticipation of the 2028 Olympics. We [Studio MMR, founded by Michelon, Casey Rehm, and Galileo Morandi in 2023] wanted to explore what those spaces could look like if they prioritized pedestrians, community gatherings, and ecological resilience. We used Cesium for real-world geospatial data, imported that into Unreal Engine, and layered Stable Diffusion and our custom-trained models to simulate new urban conditions—markets, bike lanes, green space. Then we created flythrough animations that showed these spaces evolving. We achieved what used to take months of Photoshop collages and physical models in two weeks with dynamic simulations. That speed isn’t just a technical upgrade; it changes how we communicate and collaborate.

What advice do you have for architecture’s future?

I think we’re moving toward a world where intelligence is embedded into the built environment, not just with smart sensors, but with logic, adaptability, and ecological awareness that will all be woven into the design itself. Architects can lead this transformation if we stay curious, stay critical, and stay involved. We can’t just adopt tools. We need to shape them, critique them, even misuse them creatively in the name of innovation. Automation doesn’t mean giving up control—it means redesigning the nature of control itself. In that sense, I don’t think architects are being replaced by technology. I think we’re becoming its most articulate choreographers.