EcoCircuit AI: An Autonomous Explorer for Systems – and Process-Driven Design Engineering
By Yuxin Yang, Computational Strategist; and Adam Mekies, Associate Principal
We are told all our lives, “turn off the lights,” “don’t waste water,” “use more organics,” and so on. Yet, the vast majority of us will never see where that energy comes from or where the water goes once it leaves the drain.
Environmental metabolism is not unfamiliar to urban thinkers. It can primarily be traced back to Abel Wolman, a sanitary engineer who, in 1965, published a seminal paper titled “The Metabolism of Cities.” Writing at a time when American cities were rapidly expanding and coming to terms with declining air and water quality, Wolman conceptualized cities by analogy to living organisms, analyzing their inputs (water, food, fuel, and raw materials) and outputs (solid waste, sewage, and other emissions) and quantifying the transactions between them.
Today, we still face air and water quality challenges, as well as the worsening impacts of climate change. There is a technical black box—a veil wrapped around our world—suggesting we engineers and designers and all of us citizens should be involved in change. Still, the scale, cost, complexity, and risk involved in working with physical infrastructure create astronomical barriers that often prevent problem-solving attempts, let alone allow speculative and creative freedoms.
“To lower these barriers, we developed EcoCircuit AI (www.ecocircuitai.com)—an open-source tool built to assist in systems co-optimization and strengthen the connection between digital ideas and the physical world.”
EcoCircuit AI builds on an agent capable of conceptualizing a metabolism model and a reward function for environmental relevance.
- Input and Output: Inputs refer to resources or elements introduced into the site; outputs are the resulting products.
- System: Systems include hydro, energy, biosystems, transportation, food, economic systems, etc.
- Process: Depicted by lines moving positively, representing the transformation from input to output.
- Regeneration/Feedback: Recycling or renewal of resources.
- Engineering demands: Bridging development needs and system demands.
Our research and tool position Agentic AI as a translator and communicator between Design and Engineering. Long-standing gaps exist between design-oriented thinking—which emphasizes creativity, aesthetics, and experiences—and engineering-oriented thinking, which focuses on feasibility, scalability, and technical accuracy. An AI that can “speak both languages” enables a dynamic balance, grounding imaginative ideation within the practical constraints of built system engineering. The terms “design” and “engineering” used here broadly encompass various specialized disciplines. For instance, “design” may include urban design, architecture, landscape, and interactive arts, whereas “engineering” encompasses fields like energy, water, biosystems, and robotics. The translation and collaboration between these domains, facilitated by Agentic AI—using EcoCircuit AI as a test case—aim to dissolve disciplinary boundaries, and explore and question AI’s capabilities in integration for complex environments.
Below are three case studies of EcoCircuit AI’s interpretation of Sherwood projects. Partners, collaborators, or the simply curious—please reach out and let’s make a difference in our cities’ and communities’ visionary design engineering. Please click each image to enlarge.
Dream Farm
Oceanix Busan
Woodlawn Central
