0x01 / IDENT OPERATOR → BUILDER

I build AI memory systems and agent workflows for marketing teams.

AGENT MESH · ONLINE MEMORY · SYNCED CAMPAIGN OS · LIVE POLICY · IDLE

I ran $35M+ in performance media across 20 markets with a 15-person team — then built the agents, memory and routing that operate on top. Operator-grade automation, not prompt demos.

INSPECT THE SYSTEM ▸ [email protected]

▸ live document — this page runs on the same memory graph it describes

MEMORY GRAPH · LIVE nodes 0 · edges 0
▸ idle hover a node to query memory · drag to rearrange · click a hub to jump

0x02 / READOUT OPERATOR CREDENTIALS — READ OFF THE GRAPH

0+
years in performance marketing
node/ops.tenure
$0M+
ad spend managed
node/ops.spend
0
markets operated
node/ops.geo
0
media buyers led
node/ops.team
0+
agentic systems in production
node/sys.shipped

0x03 / BRIDGE OPERATOR ⇄ BUILDER

I ran campaigns before I automated them.
That order is the moat.

● OPERATOR — LIVE EDGE

Ran the machine

Owned spend, targets and a team. Lived inside the constraints that automation has to respect — billing flows, policy walls, creative fatigue, carrier quirks.

$35M+ spend20 marketsfull funnel
● BUILDER — AUTOMATION EDGE

Builds the machine

Encodes that operating knowledge into agents, memory and workflows that run it — systems that know how the work is actually done.

agent meshmemory vaultMV3 + Claude

0x04 / CLUSTERS CAPABILITIES — EXPAND TO INSPECT SUB-NODES

Four clusters, one shared memory.

An orchestrated set of specialized agents over a shared memory vault, with a routing layer that decides who owns what. Not a single prompt — a system that holds state across tasks and across days.

strategy agentcampaign agentsocial agentcode agentmemory vaultrouting layer

Acquisition, CRO and attribution wired into one operating surface — so the system launches, measures and reallocates instead of producing slideware.

paid acquisitionCRO & LP systemsattributioncreative opsgeo scaling

A 20-rule engine that scores landing-page risk and uses Claude to generate concrete fixes — shipped as a real Chrome MV3 extension, not a checklist.

20-rule engineweighted scoringClaude fixespolicy memoryChrome MV3

Internal systems turned into shippable products with real interfaces — command centers, mobile apps, and the agentic UX patterns that make them usable.

OpenClaw consoleAI learning appagentic UXship cadence

0x05 / SYSTEMS SELECTED DEEP NODES

Shipped, not slideware.

SYS/01 · openclawPRIVATE

OpenClaw

A private multi-agent command center — one operator console commanding a mesh of agents across marketing work, with persistent memory and task routing.

agent meshmemory vaultroutingtelegram ops
SYS/02 · lp-complianceLIVE

LP Compliance Scanner

Chrome MV3 extension that scans a landing page against 20 policy rules, outputs a weighted risk score, and uses Claude to draft the fixes.

20-rule engineweighted scoreClaude-written fixesMV3
SYS/03 · learning-appSHIPPED

AI Learning App

A mobile learning product with voice-driven interaction — speech transcribed and routed into adaptive, model-generated lessons.

FlutterFirebaseOpenAIWhisper
PROMPTING→SYSTEM ARCHITECTURE DASHBOARDS→WORKFLOW EXECUTION MARKETING THEORY→OPERATOR-BUILT AUTOMATION

0x06 / EDGES WHERE I FIT — FOUR COLLABORATION PATHS

Routing four edges from one node.

EDGE/01

AI PM / Head of AI

Own the AI roadmap and the systems that deliver it, end to end — from memory architecture to shipped workflow.

EDGE/02

AI Marketing Engineer

Build the agents and automation that run marketing operations — grounded in how the work is actually done.

EDGE/03

Founder / Builder

0→1 product work where operating knowledge is the moat, not the deck.

EDGE/04

Advisor

Help teams architect their memory + agent layers without the expensive false starts.