Currently
I build practical systems for teams that need to move operations forward without waiting on a perfect software program. I stay vendor-agnostic because the right path is the one that helps the team do real work with less uncertainty and more clarity.
Why Now
AI, local models, and agent workflows are making it easier for teams to shape the systems they use without waiting on large implementation cycles.
That matters because businesses are full of people who know exactly where the friction is. The opportunity is to give them better tools, faster feedback, and safer ways to turn that knowledge into working systems.
- Less ceremony, more working tools
- Local AI where privacy and control matter
- Automation that fits how people already work
- Software that helps teams understand their own operations
Kicksights
One current example is Kicksights. It helps organizations that do not have a full internal architecture team read what is really inside a production system: automation, data paths, integration risk, and what still earns its keep.
Kicksights should help teams make clear decisions about the systems they already have. Useful systems should get easier to understand. Systems that have become drag should have a practical migration path toward tools that are easier to own and evolve.
- Map what is really in the org
- Explain architecture in plain language instead of consultancy jargon
- Identify what is valuable, risky, stale, or removable
- Support existing systems when they still earn their keep
- Define an exit path when purpose-built software is the better fit
How I Use AI
Claude
I use Claude for advising work: ideating, finding the shape of the problem, comparing paths, and turning messy context into a plan before the work is fully specified.
Codex
I use Codex when a plan needs to become files: implementation, tests, refactors, docs, and repo-native verification. It works best with a clear boundary and a way to prove the change.
Local AI and OpenClaw
I use local AI and OpenClaw for privacy, control, fast experimentation, and helping teams learn what AI can do close to their own work before anything needs production hardening.
The tradeoff
No tool should be treated like magic. Plans need implementation, implementation needs review, and local experiments need a path to something reliable if they become important.
What I Build
Business tools
Internal apps, dashboards, workflow helpers, and practical systems that replace manual work without overcomplicating the business.
Local AI workflows
OpenClaw, local models, assistants, and agent-style workflows that help people use AI while keeping more control over data and process.
Automation and integrations
The connective tissue between tools: scripts, APIs, data cleanup, notifications, handoffs, and repeatable operating paths.
Technology translation
Turning fuzzy business problems into clear options, prototypes, and implementation plans that people can actually evaluate.
Where I Am Useful
I am usually most helpful when a business has a messy process, too many disconnected tools, or a good idea that needs to become a working prototype before anyone can decide what it should become.
- Finding the repeated work hidden inside normal operations
- Turning spreadsheets, inboxes, and manual handoffs into better tools
- Pairing business context with AI in a way people can trust
- Building prototypes quickly enough to learn from real use
- Explaining tradeoffs without burying people in jargon