AI-native builder and research engineer
I work at the intersection of applied AI, data architecture, and creative tooling. My projects focus on long-context systems with persistent memory, and on interfaces that make advanced AI usable, legible, and grounded rather than abstract.
I care deeply about clarity, structure, and high-agency workflows. Most of what I build explores how to design environments—not just prompts—where humans can effectively guide and collaborate with models by scaffolding intent, shaping constraints, and aligning human judgment with machine capability across extended creative and analytical work.
Software systems that layer directives, memory, and context engineering to support long-form reasoning and creative workflows. Built around text and structured data first, with intelligent tracking of user state to enable persistent, high-agency, single-user tools that feel like a team in your pocket.
Workflow-first design for human–AI collaboration: hybrid context assembly, careful routing and orchestration across models and tools, and constraint-aware interfaces that aim for a Goldilocks zone—enough structure to prevent collapse, without overconstraining creative intent.
Multimodal pipelines for long-form writing and generative media, including diffusion-based image and video workflows. Current work centers on preventing context collapse while enabling steerability, and on formalizing these systems for broader use and collaboration.
To keep up with my latest work, follow along on Medium or Substack.