I build tools for thinking, writing, and operating with AI.

My work sits somewhere between software engineering, systems, personal knowledge management, and agentic workflows. Most of my active repositories are private because they touch local data, personal automation, operations, or early product ideas, but the recurring theme is consistent: building systems that help people make better decisions inside messy, stateful environments.

• AI-assisted operating systems for personal work, planning, memory, and follow-through
• Local-first knowledge systems that turn raw notes into durable, queryable context
• Agent workflows for inbox, calendar, repositories, research, and production
• Private tools for finance, operations, and recurring life admin
• Narrative and simulation systems where state, memory, and consequence matter
• Crypto, Bitcoin, decentralization, governance, and media analysis

I care about software that compounds.

Good tools should not just answer the current question. They should preserve useful context, reduce repeated effort, and make the next decision easier. That usually means clear source authority, durable memory, explicit provenance, and small workflows that actually survive contact with daily use.

I am especially interested in:

• agents that can work across real surfaces instead of isolated demos
• second-brain systems that maintain structure over time
• local-first tools where privacy and ownership matter
• workflow automation with human review at the right points
• interfaces that make complex state understandable without hiding the details

I spend a lot of time around:

• TypeScript, Node.js, React, Next.js
• Python for scripts, validation, and data workflows
• SQLite, Prisma, local data modeling
• Git, GitHub automation, pull request workflows
• Obsidian, Notion, markdown knowledge systems
• local automation, calendar/mail integrations, and personal tooling
• LLM-assisted development, agent memory, and structured AI workflows

I also work in crypto media and editorial systems while spending the decade prior founding a software and media house out of Manchester, UK. That background shapes how I build: I care about evidence, framing, source quality, and making complicated systems legible without flattening them into hype.

The overlap between writing and software is where a lot of my best work lives: tools that help capture context, reason through ambiguity, and turn scattered information into something useful.

Right now I am focused on building private AI systems that act less like chatbots and more like persistent collaborators:

• they remember what matters
• they know which sources are authoritative
• they separate live state from durable knowledge
• they surface the next useful action
• they keep humans in control of important decisions

Most of that work is private for now, but the north star is simple: better tools for thinking, building, and running a life with less repeated friction.