For builders, founders and operators
Multi-agent systems that are installable, observable and owned by the operator. Public proof through build logs, open tools and operating notes.
Explore AI systems →BUILDER × CLOSER × OPERATOR
I turn operator knowledge into AI systems.
Most AI portfolios show prompts. Mine shows the systems, the decisions behind them, and where they break. I build agent infrastructure for operators and sales intelligence from years inside high-ticket calls.
OPERATOR KNOWLEDGE→AI SYSTEM→PROOF
The code is not the proof. The useful artifact is. Every build here must solve a real problem, expose its limits and give you something you can inspect.
TWO COMMERCIAL DOORS · ONE PERSON
Same capability. Different buyer language.
AI systems are sold in both markets. The angle changes with the operator, not the language alone.
For founders with calls and closers
Call evidence turned into diagnosis, coaching and management signals. The commercial depth lives in Closing Code AI.
See the sales work →SELECTED BUILDS
Work before claims.
A portfolio should update every time the builder ships. These are not case-study decorations. They are inspectable proof.
QC 4.1 Light
A bilingual, portable skill that audits one sales-call transcript and returns exact evidence, breakpoint analysis and a replay plan.
Inspect build →Operator-owned agent systems
Multi-agent workflows designed around a real operating decision, with observability, boundaries and an owner.
Inspect build →Closing Code AI
The commercial depth layer for call analysis, coaching continuity, team comparison and sales direction.
Inspect build →RECENT NOTES
Field notes, not thought leadership.
What I built, what failed, the decision I made and what an operator can reuse.
OPERATING NOTE
The artifact is the portfolio
Why I am rebuilding my public work around inspectable tools, build receipts and GitHub instead of claims.
Read note →BUILD DECISION
Open source without giving away the moat
A free tool should solve a real problem. The moat lives in depth, continuity and implementation, not artificial uselessness.
Read note →