Hola!

Born at X ( Alphabet ), Everyday Robots' goal was to build a robot that could learn by itself and help anyone with ( almost ) anything. As part of the design team, I prototyped motion, light, and sound, directly on the hardware or through VR simulation to research and enhance how humans and robots could co-exist, communicate and interact.

In addition to our design prototypes, a huge effort was dedicated to developing an integrated cross-disciplinary HRI pipeline that can facilitate a smooth transition from design concept to production.

As part of our efforts in improving our robot's communication capabilities, we took on the challenge of reducing its main HRI's visual interface verbosity by delegating status messages to the robot's 'digital badge'. The digital badge is a 64x16 LED matrix, initially thought only to display the robot's identifier.

After an extensive process of categorization, semantic research, pixel puzzling, and space optimization the resulting design system had enough flexibility to display the robot's more than 50 status messages. Decluttering our main interface provided much more flexibility in our HRI efforts.

"Frags" is a physical installation for my Artist in Residence at Public Works ( San Francisco ). The installation features a set of iPads displaying the results of an evolving Stable Difusion network driven by an abstract generative input.

The convergence of all the different screens during the first stages of inference ( when pure abstract noise is displayed ) leads to the question of whether complexity is a cause or consequence of time. From this notion, a hypothetical scenario emerges where time might be actually "reversing" towards an unavoidable singularity.

The AirPods Max launch was Apple's first time launching a fully real-time WebGL-based experience. For this reason, part of the project consisted of building a workflow across multiple teams to ensure that assets displayed would meet Apple's quality standards. A Blender-based pipeline was put in place, taking care of high-density mesh ingestion, conversion to PBR workflow, and rigging for real-time. This pipeline also served to centralize asset validation by Apple's ID team as well as a "playground" for the design team to test different camera angles and lockups.

Additionally, several procedural geometries were put in place to optimize the filesize and every material was re-shaded using GLSL.

Pinterest aims to be a platform where individuals can discover inspiration for their future. That's why inspirationis core to the company’s mission. By putting each Pinner’s shifting spectrum of interests and dreams at the heart of the identity, we aimed for a living brand ecosystem that could flex into an infinite number of expressions.

In order to convey this user-centric identity, a set of tools was developed enabling anyone to create a Pinterest that would respond to their taste, creativity, and sources of inspiration.

Augmented reality experience developed for WWDC 2019 to introduce the new Mac Pro. The project was particularly tricky since Apple was releasing Reality Kit simultaneously. For this reason, our team the AR experience had also to become a demo for the framework.

The project required porting any Unity / WebGL prototypes to the proprietary platform which, being in development, required a lot of performance optimization and special shading needs. The experience contains several classic interaction gestures re-imagined to work seamlessly in AR space while providing the user with a robust 3d interaction layer.

I have been working on an several ongoing research projects involving animation and interaction in mixed reality. AR and VR have always fascinated me, and I have conducted extensive research on their capabilities and limitations across multiple platforms and devices, including AR in iOS (utilizing SceneKit, RealityKit, and Metal), Unity, and WebXR for the Oculus Quest.

All the tools I've open-sourced all the tools I've created. These include device-specific controller libraries, pipelines for in-medium edition ( webXR -> Python -> Blender -> WebXR ), and physics-based interaction models.

Motion and interaction exploration for Maslo, an AI companion. The goal of the creative tech on this project was to bring to life an abstract form that could become a visual representation of an AI. Such motion had to convey empathy and allow for a rich set of emotional responses to user input. The result is this fully parametric set of 8 concentric circles capable of traversing continuously and quantitatively the emotional spectrum.

A study by the University of British Columbia ranked Maslo's visual entity amongst the top-performing representations of an AI, along with Siri and Google assistants.