Allstar / O

Software built around language, behavior, and usable interfaces.

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.

Cute LM
SoloRoleplayer
Sifu O's Practica Dojo
Westwild

What I build

Products and systems built so far

Different builds, same pattern: text-centric software, stronger UX, local-first thinking, and a lot of hands-on iteration.

Cute LM
Local AI runtime Public product

Cute LM

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
SoloRoleplayer
Text roleplay system Public product

SoloRoleplayer//

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
Sifu O's Practica Dojo
Language learning Public portal

Sifu O's Practica Dojo

Scenario-based English and Spanish practice built around drills, pressure toggles, glue blocks, and repetition that feels like training instead of homework.

Open site
Westwild
Tabletop game world Playable beta lane

Westwild

A fast frontier tabletop setting with its own Gazette, game shell, player-facing tools, and world presentation.

Open site
Local-first runtimes and product lanes
Behavior-focused LoRA and writing adjustments
Text-centric games, systems, and tools
Proof of work real builds over inflated language

Working lanes

What WordBlock Labs is set up to do well

Local AI product lanes

Frontends and flows for people who want AI tools on their own machine, without cloud sprawl or cluttered UX.

Behavior and persona tuning

LoRA work aimed at stronger role discipline, cleaner writing style, and better persona stability.

Text-first software

Game systems, language tools, and browser interfaces where words are the product, not just decoration.

About the lab

Proof over posture

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.

Capabilities without the theater

  • Product thinking from concept to usable interface.
  • Frontend implementation with branding and interaction polish.
  • System design for local AI product lanes.
  • LLM and LoRA workflow thinking with a behavior-first mindset.
  • Clear written communication, product framing, and operational structure.

Build the specific thing

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.

Behavior is part of the product

For AI systems, the UI is only half the job. Prompt law, turn discipline, persona stability, and output quality are product work too.

Keep it understandable

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.