Accelerating real estate development and architecture
Paris Research Hackathon · 27th June 2026
The problem

Real estate is the largest asset class in the world and the slowest to move.

01

Months of feasibility, design and permits before anything is built.

02

Hard coordination between architects, engineers, consultants.

03

Projects run on debt. Every month of delay burns money.

What we do

We don't sell software. We deliver the work.

01

Feasibility studies and architectural design in days, not months.

02

Combining AI speed with in-house architect judgment.

03

No software to operate. We own the outcome.

How it works

Davis reads, thinks, delivers.

We pair AI speed with architect judgment at every step.

01

Reads

Specialized AI agents ingest the dataroom and analyze the whole project in parallel: regulations, market, site, design.

02

Thinks

Each dimension retrieved, structured and cross-referenced into what can be built.

03

Delivers

Ready-to-use outputs in days. New scenarios as fast as the first.

The technical challenge

Architecture design is not image generation.

The real problem

Buildings are structured: rooms, walls, openings, circulation, adjacencies. All bound by regulatory, site and market constraints.

Why existing tools fail

Image models break these constraints. Classical optimization does not scale to real projects.

Our approach

Specialized AI agents plus a discrete diffusion model. Outputs are structured layouts of discrete components, not pixels.

The technology
Gaudi-1Our first proprietary generative model.

Structured compositions. Constraints are taken into account during generation.


Velocity & Diversity. A wide range of distinct, buildable layouts in seconds, all respecting constraints.


Architects in the loop. Architects stay in control, from feasibility to design.

Gaudi-1 already reaches SOTA performance on established floor-plan generation benchmarks, including RPLAN and MSD, across metrics such as IoU, FID, and KID.

Gaudi-1 output
HACKATHON

The Challenge.

Mission

Build your own AI model that generates valid apartment floor plans. Given a building outline, the model should places the rooms, assigns their types and connects them into a coherent layout. Every plan must be realistic, diverse and architecturally valid.

Conditional Generation
01Framework

Diffusion or flow matching, your choice.

02Dataset

We provide access to a curated dataset of 80k single-apartment plans to train your model.

03Evaluation

Scored on held-out data across three metrics.

FID Density Coverage

Good luck!

Davis · Hackathon