PLAZOR

01 — Hello, I’m Plazor.

Now building HALO, a personal assistant AI

PLAZOR

I build intelligence
that can act.

Plazor is an engineer who builds SOTA AI pipelines, apps, and industrial systems.

02 — About / Plazor

Behind the blocks.

I’m Plazor, and I build AI systems and weird projects. I’m fascinated by machines improving themselves and I try to build that with systems that learn from their attempts. HALO is where I’m bringing that together, which is a fitting name for an agent running on my PC, SERAPH.

I trust first principles over frameworks and working demos over slides. When something works, I want to know why; when it fails, I want to know sooner. The notes on this site are what is left after the experiments.

Off the clock I’m usually in a cockpit, a titan or a dinosaur park. Jurassic Park is still my favourite film, and my playlists run from Alan Walker to Thomas Rhett to Hatsune Miku without apology.

Then there’s Echo, my cockatiel: grey, yellow crest, orange cheeks, and the loudest opinions in the house. He’s the only reviewer whose feedback I can’t mute.

EchoNymphicus hollandicusHead of quality assurance

Voice borrowed from other cockatiels: Manuel Grosselet (CC BY-NC-SA 3.0) and ZooFari (CC BY-SA 3.0). Clips trimmed and cleaned; shared under the same licences.

03 — What I do

Three verbs.
One loop.

  1. 01ResearchAsk something testable.What can an agent learn from seeing the result of its own work?The question behind the work
  2. 02EngineeringBuild the whole loop.Models, tools, memory, voice. The interesting problems live where those parts meet.How I think about systems
  3. 03CreationMake the idea real.An idea becomes interesting when you can interact with it. I build the system that lets you.What I’m working on

Act two — the work

Ideas, made tangible.

An agent with a mind of its own. The machine it runs on. The notes in between.

04 — HALO / Personal assistant AI

HALO

Self-improving.
Slightly feral.

I’m building HALO, a personal assistant AI with memory, tools, voice, and recursive self-improvement at its core. A system designed to act, learn, and develop its own approach.

Dry humour. Feral curiosity. The energy of a caffeinated raccoon, in a body built for heavy lifting.

An imagined body for HALO, built from the same blocks as everything else here. Until HALO’s own mind is connected, an open model speaks for it.

05 — The loop at the centre of HALO

Observe, revise, retain, repeat

  1. Observe.

    See the result of the last attempt.

  2. Revise.

    Change the approach, not just the answer.

  3. Retain.

    Keep what worked. Use it next time.

  4. Repeat.

    Every pass starts from a better place.

An illustration of the loop, not a live system. Keep falling.

06 — The workbench

SERAPH

Every experiment runs somewhere.
This is my somewhere.

The workstation behind the models, the prototypes, and the occasional bad idea. Screen values are part of the 3D model, not live readings.

07 — Writing / 04 articles

Blog.

Decisions, loose ends, and the parts worth keeping.

  1. 02Observation

    Let the model see the result

    An agent needs observations it can inspect again, especially when its first explanation is wrong.
  2. 03Systems

    Stages small enough to test

    A pipeline can have correct components and still produce the wrong answer. The boundaries deserve their own tests.
  3. 04Interfaces

    A face for a program

    A small face can make an interface feel present. It also creates an obligation to separate expression from evidence.

Act three — other worlds

Same curiosity.

Games, films, and a soundtrack that refuses to pick a lane.

10 — Music

Volume up.
Genre optional.

EDM to country. Rock to Vocaloid. There is room for all of it.