Working across the boundaries of academic disciplines
At Spacemaker, we are developing a game-changing AI for real estate development. In this work, we take advantage of a wide range of methods across the traditional boundary of AI. In a truly multidisciplinary approach, we combine expertise from a wide range of fields including architecture, mathematics, physics, machine learning and optimisation. We firmly believe that we can unlock tremendous value by combining state of the art methods from different technological and academic disciplines as well as by developing new methods entirely.
Breaking down real estate development to math
We are faced with a complex problem; breaking high-performance site development into its key quality components, and making site-specific design decisions based on this insight. Some of these qualities are highly quantifiable; e.g. site utilisation, noise and sun and can efficiently be quantified using tools existing tools from mathematics and physics. Other qualities are more subjective. What are the characteristics of a beautiful balcony view and how do you work to improve it for a specific project? Why do humans find certain spaces attractive and what do you do to maximize this attractiveness? For the more subjective aspects of quality, we are constantly exploring new and innovative ways to gain insight into how these aspects can be measured and optimized to improve value creation in real estate projects.
AI as a tool to drive collaboration and insights
Unlike traditional optimisation and rule-based systems, our aim is not to optimise a single objective, but to provide the user with a creative set of high-quality site proposals. At all times we are striving for the perfect balance between efficient utilization of the site and high-quality real estate. Spacemaker aims to create a more knowledge-driven real estate development process that improves value creation for developers and residents. We put a lot of effort into finding ways to facilitate collaboration and exploit synergies between AI technology, real estate developers, architects, regulators and all parties involved in the development of urban living spaces.
To do all this, we need to solve very challenging problems across a number of domains, e.g.
- Optimisation: How do we improve modelling of non-convex and non-linear properties of geometric and topological functions?
- Machine Learning: How to identify and classify preferable and unpreferable views from apartments?
- Mathematical Modeling: How can we design and combine measures that describe living qualities and at the same time works well in a machine learning and optimisation setting?
- UX and Front-End: How do we synthesize our complex AI engine into a value-creating and user-friendly product?
- Platform Engineering: How do we seamlessly scale tens of thousands of cloud instances, and transform all the massively complex data generated into an intuitive 3D-based web experience?
For the Spacemaker family, tackling these problems together is what it is all about. And this is just the beginning! We will continue to expand and improve what we do for many years to come…