Across eight years of practice, Fu has developed an architectural approach that uses emerging technology to make design, coordination, permitting, and building performance more informed.

Working across urban design, architecture, and computer science, Ye (Aaron) Fu has built experience in residential projects ranging from single-family homes to multifamily developments, as well as mixed-use, higher-education, and retail work. His practice has been centered primarily in the United States while extending to projects in China, Mexico, and the Middle East. Across those scales and contexts, he has moved from using computation mainly to generate form toward using it to organize information, evaluate performance, and manage constraints throughout the building process.

“Speed matters, but it is not the main test,” Fu says. “A useful tool should help us understand the problem more clearly and make decisions that still work when the project moves into coordination and construction.”

A Design Outlook Formed Across Scales
Fu’s academic development followed an unusual but coherent sequence: urban design in China, architecture in the United States, and later computer science. Urban 0design training taught him to understand buildings as parts of larger spatial, social, environmental, and regulatory systems. Architectural education brought that perspective down to rooms, envelopes, materials, assemblies, and individual experience.

Moving between these scales shaped his view that environmental performance is both a resource question and a matter of comfort, behavior, and everyday use. Parametric modeling, environmental simulation, BIM, and machine learning offered another way to reveal relationships that intuition alone could not always hold together. The challenge was to make those relationships useful to design rather than merely producing more output.

From Performance-Based Design to Buildable Systems
One of Fu’s earliest explorations of that challenge emerged through Arboleda, a 26-acre mixed-use community in Monterrey, Mexico, centered on multifamily housing with complementary commercial and community amenities. Fu developed parametric and environmental studies comparing orientation, massing, openings, shading, landscape, energy performance, and occupant comfort. The objective was not automation, but a more systematic way to make the consequences of design choices visible.

The process also revealed the limitations of early computational workflows. Analysis could be slow, disconnected from the primary model, or difficult to translate into a clear architectural decision. Those limitations established a principle that continued to guide Fu’s work: digital output becomes valuable only when a project team can interpret, coordinate, and act on it.

As he moved into larger projects, Fu expanded this approach from environmental analysis to fabrication logic, cost, and constructability. He developed modular facade studies to test how dimensions, repetition, and the number of unique components affected coordination and cost. For Wuhan OCT, he used parametric methods to organize variation within a repeatable exterior system, preserving an overall architectural language without solving every condition independently.

While working at CallisonRTKL, Fu applied this approach to the Xi’an Silk Road Center, developing and organizing Revit families and plugin-assisted workflows for repeated components. These tools could classify facade modules, calculate quantities, and support preliminary cost comparisons, but their task-specific nature also revealed a broader limitation: they could not interpret construction details, account for the varying difficulty of different construction methods, or understand the relationships among components beyond the specific calculation they were designed to perform. The experience led Fu to look beyond single-task automation toward tools more closely aligned with architectural reasoning—tools capable of interpreting component relationships, construction details, and practical tradeoffs rather than simply counting and classifying elements.

From Building Systems to Process Intelligence
Fu’s work on major U.S. higher-education projects broadened this understanding. At Cal Poly Humboldt, he worked across a six-building campus development program, contributing most extensively to three of the buildings and coordinating component planning, digital issue tracking, and documentation within a large multidisciplinary process. The work formed part of the University’s $458 million polytechnic transformation, an effort to expand its academic programs and reshape the campus as Northern California’s first polytechnic university. Among the program’s major projects, the $100 million Engineering & Technology Building has been described by the University as the flagship of that transformation and its first major academic infrastructure project since 2008. Working across multiple buildings made the challenge systemic: design quality depended not only on any single structure, but also on maintaining consistent component logic, information, schedules, and decisions across teams and building systems.

If Cal Poly Humboldt revealed complexity through scale, Fu’s current work at PL Design, Inc. revealed it through continuity. On residential and small-commercial projects in Southern California, he works across nearly the entire project process, including existing-condition analysis, design, zoning and code research, documentation, permitting, HOA review, consultant coordination, contractor communication, and plan-check responses. This end-to-end involvement showed that even smaller projects become information-intensive when design decisions must pass through local regulations, agency interpretations, consultant requirements, client priorities, and construction constraints. The recurring problem was no longer simply whether a design tool could perform a particular task, but whether information and decisions could remain connected, traceable, and usable as a project moved from one phase and participant to another.

Technology as an Extension of Architectural Thinking
Fu’s recent Digital Twin AI-Powered Intelligent Building Platform brings these concerns into the operation of buildings. Developed for an international design competition, the proposal integrates building models, operational data, energy information, and spatial context within a unified interface for campus managers and other decision-makers. Fu developed the platform’s architectural framework and interaction logic, organizing information in relation to spaces and building systems rather than presenting operational data as isolated charts.

The platform is designed to support energy monitoring, facility management, spatial use, maintenance, and long-term planning, with AI identifying patterns while keeping professional expertise in the decision-making process. It marks an important shift in Fu’s trajectory: his early computational studies evaluated buildings before construction, while the Digital Twin proposal extends architectural intelligence into occupancy and operation.

“Construction may be complete, but a building continues to evolve through use,” Fu says. “Its performance depends on how it is operated, maintained, and occupied over time.”

Toward Responsible Architectural Intelligence
Fu’s practice is increasingly focused on bringing architectural reasoning and computational intelligence together. Graduate study in computer science has deepened his technical foundation while also making him more critical of claims that AI automatically improves efficiency. A system may generate alternatives quickly yet create additional work later if it ignores dimensions, codes, structural logic, budgets, client requirements, or constructability.

This critical perspective also shapes his research at the beginning of the building life cycle. His work on residential zoning feasibility explores how requirements can be translated into structured, machine-readable information while keeping each conclusion connected to the governing rule. More broadly, he is interested in transparent tools that can clarify constraints, compare alternatives, preserve decision history, and improve coordination across project phases.

“The goal is not to remove professional judgment from the process,” Fu says. “It is to make the relevant constraints visible, show how a conclusion was reached, and give architects better information for making decisions.”

Looking ahead, Fu plans to develop integrated building-intelligence systems connecting design, permitting, construction coordination, and operation. Such systems could bring regulation, energy performance, cost, and long-term use into a continuous framework while preserving the reasoning behind each recommendation. For Fu, the future of architecture is not a choice between human and machine intelligence, but a question of how responsibly the relationship between them is designed.