My Recent Projects
Blueberry AI - Founder
Blueberry AI is a new type of nutritional tracking app, built from the ground up for the AI Age. It features the most advanced AI Nutrition Engine on the market.
I am the sole designer, engineer, and founder. Blueberry is currently running a private waitlist for TestFlight users as we are in the final weeks of development.
Beta opens
Sept 2025
Founded in
Feb 2025
Blueberry AI - Founder
Blueberry AI is a new type of nutritional tracking app, built from the ground up for the AI Age. It features the most advanced AI Nutrition Engine on the market.
I am the sole designer, engineer, and founder. Blueberry is currently running a private waitlist for TestFlight users as we are in the final weeks of development.
Beta opens
Sep '25
Founded in
Feb '25
Blueberry AI - Founder
Blueberry AI is a new type of nutritional tracking app, built from the ground up for the AI Age. It features the most advanced AI Nutrition Engine on the market.
I am the sole designer, engineer, and founder. Blueberry is currently running a private waitlist for TestFlight users as we are in the final weeks of development.
Beta opens
Sept 2025
Founded in
Feb 2025
Digital Twin
Fine-tuning a model to contain my memories, personality, and habits. Data from my iPhone, tuned using LORA and local compute (2 Mac Studios), and run on a 7b model (Mistral) that is capable of running inference on edge. Learned everything on the fly.
Project length
2 weeks
Training examples
100k
Digital Twin
Fine-tuning a model to contain my memories, personality, and habits. Data from my iPhone, trained using local compute (2 Mac Studios), and run on a 7b model (Mistral) that is capable of running inference on edge. Learned everything on the fly.
Project length
2 weeks
Training examples
100k
Digital Twin
Fine-tuning a model to contain my memories, personality, and habits. Data from my iPhone, tuned using LORA and local compute (2 Mac Studios), and run on a 7b model (Mistral) that is capable of running inference on edge. Learned everything on the fly.
Project length
2 weeks
Training examples
100k
Generative Computing
"The last program ever written will be that which creates other programs"
The second iteration of the Generative UI prototype, this breakthrough garnered widespread excitement throughout Apple, launching significant investment into the space. I lead development of Generative UI at Apple, as the leading expert in the subject.
Model used
GPT-4o
Presented to
300 ppl
Generative Computing
"The last program ever written will be that which creates other programs"
The second iteration of the Generative UI prototype, this breakthrough garnered widespread excitement throughout Apple, launching significant investment into the space. I lead development of Generative UI at Apple, as the leading expert in the subject.
Model used
GPT-4o
Presented to
300 ppl
Generative Computing
"The last program ever written will be that which creates other programs"
The second iteration of the Generative UI prototype, this breakthrough garnered widespread excitement throughout Apple, launching significant investment into the space. I lead development of Generative UI at Apple, as the leading expert in the subject.
Model used
GPT-4o
Presented to
300 ppl
Personal Deep Research
Imagine “LLM Deep Research,” but pointed inward: the model searches your iPhone’s personal data to answer questions about you.
By pairing on-device retrieval (RAG) with an LLM, this prototype resurrects personal memories on demand and hints at the future of truly personal assistants.
Retrieval system
RAG
Model params
14b
Personal Deep Research
Imagine “LLM Deep Research,” but pointed inward: the model searches your iPhone’s personal data to answer questions about you.
By pairing on-device retrieval (RAG) with an LLM, this prototype resurrects personal memories on demand and hints at the future of truly personal assistants.
Retrieval system
RAG
Model params
14b
Personal Deep Research
Imagine “LLM Deep Research,” but pointed inward: the model searches your iPhone’s personal data to answer questions about you.
By pairing on-device retrieval (RAG) with an LLM, this prototype resurrects personal memories on demand and hints at the future of truly personal assistants.
Retrieval system
RAG
Model params
14b
Operator for iPhone
Computer-use agent for iOS, similar to OpenAI's Operator, with a virtualized machine capable of running an infinite number of iPhones.
Tasks are split into human understandable chunks, and spread across a dozen iPhones, each with an agent in control of the UI.
Model
Claude
Virtual machines
12 iPhones
Operator for iPhone
Computer-use agent for iOS, similar to OpenAI's Operator, with a virtualized machine capable of running an infinite number of iPhones.
Tasks are split into human understandable chunks, and spread across a dozen iPhones, each with an agent in control of the UI.
Model
Claude
Virtual iPhones
12
Operator for iPhone
Computer-use agent for iOS, similar to OpenAI's Operator, with a virtualized machine capable of running an infinite number of iPhones.
Tasks are split into human understandable chunks, and spread across a dozen iPhones, each with an agent in control of the UI.
Model
Claude
Virtual machines
12 iPhones
Copresence
Engineered AR prototypes for Apple Vision Pro’s FaceTime telepresence features, securing executive approval during product development. These features played a key role in the creation of FaceTime marketing materials for this new category of computing.
Final product
Facetime
Platform
visionOS
Copresence
Engineered AR prototypes for Apple Vision Pro’s FaceTime telepresence features, securing executive approval during product development. These features played a key role in the creation of FaceTime marketing materials for this new category of computing.
Final product
FaceTime
Platform
visionOS
Copresence
Engineered AR prototypes for Apple Vision Pro’s FaceTime telepresence features, securing executive approval during product development. These features played a key role in the creation of FaceTime marketing materials for this new category of computing.
Final product
Facetime
Platform
visionOS
Dynamic UI
Develop an LLM-based system capable of creating iOS and visionOS interfaces instantly. This enables immediate use cases in Accessibility, localization, and personalized workflows. Originally aimed for a distant iteration of visionOS, it emphasizes the need for highly personalized software in a future with ubiquitous AR & AI devices.
Model
GPT-3.5
UI generation in
20 sec
Dynamic UI
Develop an LLM-based system capable of creating iOS and visionOS interfaces instantly. This enables immediate use cases in Accessibility, localization, and personalized workflows. Originally aimed for a distant iteration of visionOS, it emphasizes the need for highly personalized software in a future with ubiquitous AR & AI devices.
Model
GPT-3.5
UI generation in
20 sec
Dynamic UI
Develop an LLM-based system capable of creating iOS and visionOS interfaces instantly. This enables immediate use cases in Accessibility, localization, and personalized workflows. Originally aimed for a distant iteration of visionOS, it emphasizes the need for highly personalized software in a future with ubiquitous AR & AI devices.
Model
GPT-3.5
UI generation in
20 sec
Magic Mocap
From RGB vide recording to professional Motion Capture — harnessing Apple's ARKit platform to put a full motion capture studio in your pocket.
Framework
ARKit
Reduce cost by
-90%
Magic Mocap
From RGB vide recording to professional Motion Capture — harnessing Apple's ARKit platform to put a full motion capture studio in your pocket.
Framework
ARKit
Reduce cost by
-90%
Magic Mocap
From RGB vide recording to professional Motion Capture — harnessing Apple's ARKit platform to put a full motion capture studio in your pocket.
Framework
ARKit
Reduce cost by
-90%
Portal Cube
Three day sprint to discover a new paradigm for frictionless window management in Mixed Reality. Inspired by non-euclidian geometry in video games, this prototype brings forth a new method for saving space when handling multiple windows.
Project length
3 days
Target platform
visionOS
Portal Cube
Three day sprint to discover a new paradigm for frictionless window management in Mixed Reality. Inspired by non-euclidian geometry in video games, this prototype brings forth a new method for saving space when handling multiple windows.
Project length
3 days
Target platform
visionOS
Portal Cube
Three day sprint to discover a new paradigm for frictionless window management in Mixed Reality. Inspired by non-euclidian geometry in video games, this prototype brings forth a new method for saving space when handling multiple windows.
Project length
3 days
Target platform
visionOS
Vision Pro Peripherals
A personal project looking into possible peripherals for the Apple Vision Pro. Hardware peripherals would enable custom workflows, especially for pro and power users.
Prototyping tool
Lego
Target user
Pro
Vision Pro Peripherals
A personal project looking into possible peripherals for the Apple Vision Pro. Hardware peripherals would enable custom workflows, especially for pro and power users.
Prototyping tool
Lego
Target user
Pro
Vision Pro Peripherals
A personal project looking into possible peripherals for the Apple Vision Pro. Hardware peripherals would enable custom workflows, especially for pro and power users.
Prototyping tool
Lego
Target user
Pro
Projects from High School
Here you will find some previous work in hardware, software, and product design from my High School & College years. Since childhood, I have remained intensely passionate and curious about the future of computing.
Published first app
13 yrs old
Total downloads
25k+
Projects from High School
Here you will find some previous work in hardware, software, and product design from my High School & College years. Since childhood, I have remained intensely passionate and curious about the future of computing.
Published first app
13 yrs old
Total downloads
25k+
Projects from High School
Here you will find some previous work in hardware, software, and product design from my High School & College years. Since childhood, I have remained intensely passionate and curious about the future of computing.
Published first app
13 yrs old
Total downloads
25k+