Local AI product lanes
Frontends and flows for people who want AI tools on their own machine, without cloud sprawl or cluttered UX.
Allstar / O
WordBlock Labs is my proof-of-work site and learning laboratory. I am a former Fortune 30 senior-level executive who saw the AI shift coming, pivoted hard, and taught myself this stack so I could keep moving with the industry instead of watching it happen from the sidelines.
Most of what I build lives close to words: local AI runtimes, roleplay systems, language practice tools, tabletop game interfaces, and training workflows for better model behavior. The throughline is simple: make the software clearer, more disciplined, and more useful than the generic version.
What I build
Different builds, same pattern: text-centric software, stronger UX, local-first thinking, and a lot of hands-on iteration.
A privacy-first local language model launcher for people who want models running from their own machine with a calmer, cleaner setup path.
Open site
An AI solo roleplayer experience built for local play, memory continuity, and cleaner turn ownership, with a strong focus on reducing head-hopping and persona drift.
Open site
Scenario-based English and Spanish practice built around drills, pressure toggles, glue blocks, and repetition that feels like training instead of homework.
Open site
A fast frontier tabletop setting with its own Gazette, game shell, player-facing tools, and world presentation.
Open siteWorking lanes
Frontends and flows for people who want AI tools on their own machine, without cloud sprawl or cluttered UX.
LoRA work aimed at stronger role discipline, cleaner writing style, and better persona stability.
Game systems, language tools, and browser interfaces where words are the product, not just decoration.
About the lab
This site is meant to show what I can build, how I think, and how I execute. I am not trying to inflate the story. The work should do that part for me.
I come from a senior corporate background, but this chapter of my life has been about getting hands-on again: shipping, testing, learning the stack, and building things that work.
I prefer purpose-built systems over broad vague ones. A roleplay shell should feel like a roleplay shell. A local runtime should feel calm and direct. A language tool should feel like practice.
For AI systems, the UI is only half the job. Prompt law, turn discipline, persona stability, and output quality are product work too.
I like software that can explain itself. Clean layout, plain language, honest constraints, and a setup path that does not waste the user's time.