Yekta Ipek

“Computational design allows designers to address different concerns in the very early days of design.”

Yekta Ipek is a computational design expert and associate partner of Pilbrow & Partners. He told us how computational design is currently shaping the architectural design process.

Can you tell us a bit about your professional background?

I started my undergraduate studies with a bachelor’s degree in civil engineering at Istanbul Technical University. In my first year, I decided to start double major degree in architecture along with civil engineering degree. After finishing both degrees, I finished my master’s degree in Architectural Design Computation again at ITU. During my master’s degree, I was working for a local company, Uras X Dilekci Archiects, where I found the opportunity to practice my theoretical courses in professional works. It was challenging to find the solutions to the real-world problems in architecture by using the knowledge that I was gaining from the courses. During my graduate studies, I joined to the University of Salford (UoS) in Manchester as an exchange researcher to complete my thesis on computational approaches in architecture. I spent 6 months in UoS and had the chance to get exposure on the perspective of “digital” concepts in the UK, especially on Building information Modelling (BIM). After completing my research in the UK, I went back to Istanbul and finished my graduate study. I worked for Henning Larsen Architects and DOME + Partners until 2014. In 2014, I moved to London and worked for several international and national companies in London. I am now working as an Associate Partner at Pilbrow & Partners, where I joined the practice in 2017.

What is your role in Pilbrow and Partners?

My role in Pilbrow & Partners is integrating the computational workflows into daily designer’s routine and to use the tools available in the industry in an efficient way to make the process productive for the company by avoiding the redundant work.

Which projects are you currently working on?

As P&P, we mainly work on national projects. We also have some international works, but most of our projects in London. I mostly work on master plans and large-scale mixed-use developments.

How do you define parametric design and computational design? What are the differences?

It is hard to differentiate the two concepts for me, but I would say parametric design includes parameters you produce the project/process with, whereas computational design includes the methods that you produce the project/process with computers. To me, we are in an era where most of the concepts got integrated and become as a holistic methodology for us when we design our products. I would define computational design as a medium we can use during the design process, and parametric design is an approach to the design process. Computational design mediums give us the ability of processing the parameters in a complex and efficient way and creating the output in a powerful manner, avoiding the exhaustive practice.

Computational design is becoming a highly complex sub-field in architecture that requires specialists like you. On the other hand, a project relies on the expertise of many specialised consultants. What do you think about the extreme branching of the architecture into sub-fields? How does it affect the design process and the principals of practices?

Yes, it is correct. It requires a lot of computer knowledge like programming, scripting, modelling etc…, so you need the special knowledge to integrate it into design process. But this statement is valid for now. I believe it will be invalidated in the coming years. We will see that there will be lots of tools that do not require the special knowledge to use and integrated into our design workflows. As the concept gains more significance, designers get familiar with the required knowledge, and the concept becomes as one of the core components of the design process, like today’s simple CAD software.

As designers, I believe, we need to have the holistic understanding of the projects such as technical, social, environmental, economic aspects, etc…. That is why we work with specialised consultants to consider the different aspects when we design. When we collaborate, especially in the early days of design, we generally struggle to provide the information that they need, as they assess the design works in a very detailed way. Since the design is not a linear process and requires lots of iteration; I believe, the profession should manage to integrate computational design tools to take sensible decisions when they test different paths to the end-product from the very beginning of the process. Of course, we need the specialised consultants, but it is vital for practices to make the computational design one of their core components in their workflows to meet the requirement of today’s fast-paced environment.

Computational design allows designers to address different concerns in the very early days of design, so they can create more favourable and the right products for the society.

How do you do this? Can you give a more concrete example where your expertise helped the design process or concept?

We are using the computational design methods in different aspects of the projects and at the different stages as we need to take the decisions along the way. For example, when taking massing decisions at the early stages for the projects in the historical context, we need to be very careful about its visibility from the key viewpoints. What we do is that we take the key points and calculate the invisibility envelope against the context – using detailed context models so we can get as accurate as possible. Then we start working within that envelope to mitigate townscape problems later the design work. Imagine you have 20 key viewpoints to test. With the help of the computation capabilities, you can quickly compute the invisible massing form based on these 20 points. This gives a descriptive feedback to the design process. Another good example would be the performance aspect of the project such as structural, environmental, etc. When designing a structural shell, we explore different structural articulations to address the aesthetic concerns. It is essential that each iteration should be structurally conscious so the design progress is in the right direction. With the generative methods, we can implement algorithms that your structure articulation forms itself based on the structural performance on the given conditions. It creates the opportunity to integrate the structural performance input into design from the beginning. I hope these two examples give an idea how I am implementing these methods into design process.

How do you work with your specialised consultants?

In the practice, we aim to integrate our computational tools from the very early days so we can take sensible decisions as we progress our design works in order to avoid the further risks along the way. We, then, share our design with the consultants to get a comprehensive perspective from them.

Is synchronisation of information and tools still a problem while working with external consultants or is this a resolved issue?

I do not think it is a significant problem as it was before. When we use same tools/environments with our external consultants, we exchange the information quickly and smoothly without any loss. On the contrary, with the consultants that are using different tools, which is the case generally as they are using more specialised tools for their work, we exchange data using the exchange mediums. Of course, it takes more time and effort to create these common grounds to exchange the information but it is possible, especially with the help of more developed software.

We are becoming used to AI in many fields. Robots, driverless cars, automated delivery drones are not part of a utopia any more. Shopping, manufacturing and even farming is becoming more automated and simple. On the other hand, architecture is becoming a field that relies more and more on the human workforce with lots of decision makers and endless paperwork and regulations. Yet we can’t avoid burning or leaking buildings. Why don’t we see the same pace of development in architecture as we see in other fields of life?

We have now lots of developments in the industry compared to the last years. Yes, I agree we could not have achieved the same amount of development as it has been achieved in the other industries. I think the main reason is that there is a lack of resource allocation for R&D in the practices. That is why we are short of technologies in the industry in general terms. However, because of the nature of the design, I believe it is very complex to imitate designers’ intentions when it comes to automation and AI side. We can replicate the routines that designers follow, but it is difficult to depict the heuristic side of designers in the process. There are tools that help the automation of the design and remove the manual labour; however, there is a great potential to extend it further. I think with the help of computational design especially BIM we will automate the paperwork labour in the construction process and eliminate the manual workforce in the coming years. But I am not sure about the automation in the design aspect, as I explained earlier. It will require more development in AI from my point of view.