In Weitzman’s newest Master of Design program, students are using AI to integrate design and finance—and developers are taking note.
"Through computation, we are linking geometric considerations with financial considerations," says MSD-PDD associate director Richard Garber (left). Summer session faculty member Kiran Chokshi spoke to students and guest critics at the final review. (Photo Kait Privitera)
Over lunch at a diner, a group of developers chat about the possibility of a new housing project in Philadelphia’s Fishtown neighborhood. They scribble a flurry of numbers to estimate its viability: lot size, number of units, cost of construction.
'Like an architect making a napkin sketch, these real estate professionals ask, “Does it pencil?”
“Penciling” is just one of the new vocabulary words for students in the first cohort to pursue a Master of Science in Design: Property Development and Design (MSD-PDD). This summer, they bridged the gap between architecture and finance by using AI (not actual pencils) to integrate financial planning and zoning regulations into the design process.
Architect Richard Garber, the MSD-PDD program’s associate director and a senior lecturer in the Department of Architecture, explains that the use of AI to test real-world scenarios around property development is an evolution of how architects use advanced large language models (LLMs). Rather than simply produce a stream of images, they are developing AI-assisted apps that work extremely quickly between traditional programs like Excel and Rhino to optimize the entire housing design and development process for a standard 5-over-1 mid-rise development in Philadelphia, the ubiquitous housing type with retail on the ground floor and several floors of housing above.
“AI allows us to use real 2D and 3D data to better design each site for what real estate people call ‘highest and best use,’” Garber says, “For example, formerly industrial sites are being rezoned to foster more productive kinds of development like mixed-use with housing and retail.”
The MSD-PDD summer studio was led by Garber, who is also a principal at GRO Architects, with architect Shajay Bhooshan of Zaha Hadid Architects, and Kiran Chokshi, a Wharton graduate and founder of Forkast, an AI finance startup. Both have extensive experience using AI to supercharge decision making and the design process.
In some instances, students with little to no design experience could do in an afternoon what would typically take three consultants weeks writing scripts in Grasshopper, the Rhino plug-in typically used to create parametric models, automate design tasks, and generate complex geometries. “These students have not experienced property design and development from the perspective of finance,” Chokshi says. “The apps they developed have allowed them to connect zoning, geometry, and financial feasibility like never before.” Bhooshan helped the students produce .3dm Rhino files based on AI analysis of inputs such as zoning envelopes, facade generators, and quantity takeoffs.
Xiaoyang Qiu’s project is a mid-rise residential building near Rittenhouse Square for which she built custom web app prototypes.
For the studio, each student chose a site in Philadelphia. Xiao Yang Qiu’s project is a mid-rise residential building near Rittenhouse Square. She built custom web app prototypes to test and deploy solutions, including a building envelope tool that uses Philadelphia’s actual affordable-housing Floor-Area-Ratio (FAR) bonus to recalculate the floor count as the building expands in design.
“Architects naturally push aesthetic and structural boundaries, but real-world execution is always shaped by practical constraints,” says Qiu, who came to Penn from China Academy of Art, where she majored in Architecture Design. “This studio taught me how to translate design concepts into the distinct "languages" of three key stakeholders—the municipality, the developer, and the community—enabling smoother communication and a more viable path to project delivery.”
Qiu made two different custom AI apps for the design. The “Structural Load Visualizer” imported beam and column centerlines, to quickly visualize load distribution and simulate how added shear walls or structural core would change structural needs. The “Hard Cost Breakdown Calculator” broke down hard costs item by item, allowing Qiu to estimate how architectural elements would save material and reduce expenditures. Qui also used AI to solve daylighting issues including a solar analysis to guide varying balcony depths.
Hunter Gutt’s proposal at 451-63 North 10th Street is a typical 5-over-1 mixed-use residential building with a makerspace on each residential level and ground floor art retail and exhibit space. “My goal was to take this common type of development proposal and elevate it to a higher level of design and thought than the bland, placeless 5-over-1's we see popping up everywhere,” says Gutt, who previously received his MArch from Weitzman in 2026 after completing an Architectural Studies undergraduate degree at College of the Holy Cross in 2023.
Gutt says that he appreciated the real-world perspective, which was a change from traditional theory-based studios. “In a MSD-PDD studio, I had to defend an elaborate glazing and facade system with realistic market price assumptions while maintaining a coherent argument for fabrication methods,” he says.
“I had to defend an elaborate glazing and facade system with realistic market price assumptions while maintaining a coherent argument for fabrication methods,” says MSD-AAD student Hunter Gutt, who used OpenAI, Claude, and Gemini; along with the PennChat AI suite and 3D printing to design a proposed new housing development in Philadelphia.
To differentiate the development from run-of-the-mill offerings, Gutt included a solar responsive facade paneling system and a 3D-printed concrete first floor podium structure. He used AI, including OpenAI, Claude, and Gemini; along with the PennChat AI suite access as an assistant. “I used a workflow that weaved in and out of manual research and modeling, coding apps and rhino python scripts, and parametrically connecting the produced meta-data to an active pro forma that would adjust along with the design iteration process,” he says. (The pro forma is the financial document that projects a property’s cash flow, showing income and expenses.)
The MSD-PDD faculty are enthusiastic about bringing together Wharton and Weitzman skills, and the potential for agentic, code-writing AI. By searching for the “highest and best use” for each site, architects can speak the language of developers and expand their own agency outside the traditional design services role that comes after the financials are set. Bhooshan explains, “We tried to push the limits of the integration of domain expertise (spatial AI, housing specifics and financial intelligence) and custom site and development parameters.”
The direct engagement with finance that MSD-PDD students have had could give architects new ways to contribute to the housing crisis. “Through computation, we are linking geometric considerations with financial considerations,” Garber says. “So now architects can engage larger ideas about property development and working with property owners.”
AJ Pires (MArch’06), a Weitzman alum and the president of Brooklyn-based Alloy Development, sees the same potential. “It was great to see students grappling with the very real economic, construction market and regulatory constraints that inform our built environment. Their ability to understand and navigate these systems will afford them more access and agency as they go out into the world,” he says.
For all their promise, the AI-based tools aren’t expected to replace architects any time soon. “It is about giving people more opportunities and capabilities,” Garber says. "Every good prompt had three parts: a goal, a rule, and a test it had to pass. You can't write that unless you already know zoning and you already know a pro forma."
"Students didn't ask AI to design a building,” Chokshi says. “They asked [large language models] to build the instruments that tell them whether a building is worth designing.”
View Slideshow
