Skip to content

Founding Cohort Demo

Gigi Palma

Founding Cohort

Gp's ZeroClaw Lab

A personal agent experiment lab documenting Nova's design, weekly lessons, and an early real estate analyzer showcase.

Project Snapshot

Builder

Gigi Palma

Strategic Program & Transformation Leader | Scaling Global Teams, Risk & Operations | PMP | Oxford MSc

Lead Agent

Nova

The architecture section of the live demo is centered on Nova and the agent identity behind the lab.

Showcase

Real Estate Analyzer

An early public project used to demonstrate how the lab turns agent experimentation into a concrete use case.

What The Demo Shows

The live site is positioned as a personal experiment lab rather than a polished product launch.

It explicitly documents three things:

  • Nova agent architecture
  • weekly learnings captured in a journal
  • a real estate analyzer showcase project

That framing matters. The project is not pretending that the interesting part is only the finished output. It treats the system design and the learning process as part of the asset.

Structure

The published demo is organised into three main sections:

Architecture

The architecture section is built around Nova and breaks the agent down into:

  • identity
  • soul
  • skills
  • user
  • style

This suggests a deliberate effort to define the agent beyond functionality alone, which is a useful pattern for any member building an owned system with a recognisable behaviour and role.

Journal

The journal currently exposes weekly learnings, including a Week 3 Lessons entry. That makes the lab a working log of iteration rather than a static portfolio page.

Showcase

The showcase currently highlights a Real Estate Analyzer project. In the context of the lab, this functions as the proof point for how the underlying agent design can be applied to a concrete domain problem.

Why It Matters

Gp's ZeroClaw Lab is a useful founding cohort example because it documents an agent system at the right level of fidelity:

  • the agent identity is explicit
  • the build process is visible
  • the learning loop is part of the output
  • there is already a clear application example

That is the right direction for a member lab. It turns experimentation into an owned body of work instead of leaving it as a set of disconnected prompts or notes.

Visit the Demo

Closing Thought

Gigi's project is still early, but the structure is correct: define the agent, document the lessons, and connect the architecture to a practical showcase. That is the kind of build record that scales as the system becomes more capable.