Intelligence Lab

A neural reef for adaptive AI.

The patterns and primitives I keep reaching for. Designed to be measured, not just demoed.

Live
Sparring Agents

Persona-driven LLM agents — scoped, evaluated, re-rubricked weekly.

Production
Conversational Simulations

Branching narratives with adaptive coaching. Intent, not keywords.

Beta
Workflow Orchestration

Trigger → diagnose → reinforce → measure → reroute.

Open framework
Prompt Systems

Reusable scaffolding, persona libraries, rubrics — versioned like code.

Prototype
Performance Support Agents

In-flow assistants surfacing the right signal at the right moment.

Core
Eval & Guardrails

Rubric-driven evaluator harness, drift detection, SME-in-the-loop.

L2Inside the system

Versioned like code. Evaluated like product.

persona.objection-handler.v3.ts
{
  "persona": "skeptical-VP-Eng",
  "context": ["scale-up SaaS", "post-Series-B"],
  "guardrails": {
    "no-pricing-claims": true,
    "no-roadmap-promises": true
  },
  "rubric": [
    "names the real risk",
    "offers proof not theater",
    "ends with a forward step"
  ],
  "eval": "human + LLM-judge"
}
Live eval harness
Conversation scoring
Discovery depth84
Objection acknowledgment91
Forward motion76
Tone calibration88
Open simulation

Step into the interactive sales pitch activity.

A guided, interactive experience that helps learners structure and deliver an effective sales pitch — discovery, value framing, and forward motion.