I build systems
that think — and
interfaces for humans
to think with them

Engineering leader and ML researcher working across every layer of the stack. From systems programming and native mobile to large language models and autonomous agents — I architect products, lead teams, and publish research at the frontier of applied AI.

View Work Hugging Face

What I Do

Three modes of building

I think in systems, ship as a product engineer, and research at the frontier of machine intelligence.

Applied AI & Agents

Building intelligent systems that reason, plan, and act. I fine-tune large language models, design agentic architectures, and research approaches to long-horizon reasoning and continuous self-improvement. Published 10+ models and datasets on Hugging Face, including a 21B parameter fine-tune.

Product Engineering

End-to-end product thinking with deep technical execution. I architect and ship full-stack web applications, native iOS and Android apps, and watchOS experiences — always with a systems mindset. Seven apps on the App Store and counting.

Human-Computer Interfaces

Designing the surfaces where humans and intelligent systems meet. From pixel-perfect mobile UIs to voice interfaces to embodied intelligence — I'm interested in every modality through which people interact with computation.

Research Interests

What I think about

The problems I keep returning to — at work, in open source, and in the margins of my notebooks.

001

Long-Horizon Reasoning

How do we build agents that don't just react, but plan? I'm interested in architectures that enable multi-step reasoning, goal decomposition, and execution across extended time horizons — agents that can hold a thread for hours, not just tokens.

002

World Models

Intelligence requires a model of reality. I explore how to ground language models in structured representations of the world — spatial, temporal, causal — so they can predict, simulate, and reason about consequences before acting.

003

Continuous Self-Improvement

Systems that learn from their own experience. I research approaches to self-play, reflection, and iterative refinement — how an agent can evaluate its own outputs, identify failures, and improve without human intervention.

004

Embodied Intelligence

The full spectrum of human-computer interaction — from touch screens to voice to physical embodiment. I'm drawn to the question of how intelligent systems should present themselves to humans, and how that interface shapes what's possible.

Models & Publications

Open research

Models, datasets, and technical writing published on Hugging Face.

Model

GPT-OSS-20B-SemanticWiki (GGUF)

21B parameter model fine-tuned for semantic code understanding

318 downloads

Dataset

SemanticWiki Data

Structured dataset for semantic code analysis and documentation

1.2K downloads 41 likes

Article

Unlocking Agentic RL Training

Exploring reinforcement learning for autonomous agent training

Technical deep-dive

Model

FunctionGemma 270M Mobile Actions

Compact model for on-device function calling and mobile actions

Mobile-optimized

Featured Work

Things I've shipped

Products, tools, and experiments — from AI agents to watchOS apps.

Latest Launch

Daily Quest

Transform your daily tasks into epic adventures. A gamified productivity app that turns your to-do list into quests, complete with experience points and level progression.

Swift SwiftUI CloudKit App Store

Try Daily Quest →

⚔️

Open Source