HOME
Software - TouchDesigner, TouchOSC, Arduino IDE, Polycam, Blender, Photoshop, Illustrator, After Effects
I love what I do, and I want to do as much of it as possible. Living in a warehouse in North London, surrounded by creatives of all kinds, has given me the freedom to dive into all sorts of experiments, especally as the “Renter Frendly” mentality is no longer a limitation.
Shown here are a few of those explorations, from turning my room into an interactive projection-mapped canvas to creating custom projection, linked lighting and projection systems for the warehouse parties we host each year, and clothes I design and make on my Singer hand driven sewing machine from the 1920s.
Software - Unity , Blender, Illustrator
Possibly the start of my love for boid systems. This project was created during the Bartlett Introductory Workshop, taught by the brilliant Elena Polychronaki of Uncrated Limbo.
coolrtitle_ was built in just two weeks by fellow student, and genius, Elena Petrova and myself. It's a small DJ-simulator experience where an audio-reactive flock of boids dances in front of the player.
The exported Unity project could also generate souvenir 3D models from the boids' improvised choreography, using a marching cubes process to turn their paths into exportable digital objects.
These are from a little charity art auction in Hackney: three posters and three postcard designs, each with its own augmented reality overlay.and they still work! slide through and find those qr codes to the left and give it a go.
Software - 8th wall (RIP), Blender, zBrush, After Effects
Everything to the right is the work of my wonderful students. Please get in touch if you'd like to be connected with any of them.
For the past three years, I've taught a two-week Bartlett introductory workshop for incoming MArch students. Each year has explored a different mix of tools and techniques, from augmented reality development, AI image generation, and simulation workflows, to storytelling, organic modelling, animation, and compositing.
Last year, we brought many of those strands together, combining storytelling, modelling, animation, and AR development into a single workshop. It's always incredible to see what people can achieve in just two weeks when they're given the tools and support to explore.
ChatGPT launched about a month after I started my master's. Rather than having it do the project for us, my team and I decided to give LLMs bodies and scatter them around London.
The project explored how people might interact with autonomous digital forms in an urban AR context. I spent a lot of time prototyping in Unity, using boids and marching cubes to create interactive creatures. Their appearance was generated from user-captured images using CLIP Interrogator and Stable Diffusion, while internet data from their location informed their personalities.
The also addresses open-source versus proprietary technologies and the future of cyber-physical inhabition. I’ll leave it there but feel free to ask me more about it.
Just before the rise of diffusion models, StyleGAN was a common approach to image generation.
Unlike diffusion models, which iteratively refine noise into images, StyleGAN maps a latent vector through a learned space to generate outputs. It’s trained using a generator–discriminator setup, where the discriminator learns to distinguish real images from generated ones, and the generator improves by learning to fool it.
For this project, I trained a model on hundreds of images of seahorses, running it on a 2080 GPU for around two days. The result is a set of interpolations through the model’s latent space, showing gradual transitions between generated forms.
At the time, the work explored camouflage and visibility, seahorses blending into their environment, and the system’s attempt to learn what it means to “see” them.