Intelligence should earn your trust.

I’m Leandro Fajardo. I build and explore systems that connect AI intent to accountable action.

Apple Valley, CA · Independent builder · 2026

Explore the engineering   /   See the bigger vision   /   Build with me

An agent operating a device. A printer producing a part. An AI helping create a work. A retrofit changing a space.

Each raises the same questions:

Who authorized the change? What are the limits? What proves the result?

That is the thread connecting my engineering experiments and venture concepts. I call it Trusted Transformation Infrastructure: clear authority, bounded execution, and outcomes people can inspect.

Build with boundaries. Verify outcomes.

My engineering work explores AI agents, native Apple workflows, and developer automation. The focus is the path from a request to a useful, checkable result.

Design priorities: private processing where practical, explicit permissions, native interfaces, observable execution, and recovery when something fails.

These repositories are a working lab. Check each project’s documentation for its implementation and current status.

The same control model extends beyond software. Three original venture concepts explore what happens when physical and creative transformations carry clear rules and evidence.

Modular local production hubs for customized, low-volume goods.

First proposed test: two qualified local printers, one replacement part. Compare lead time, quality, landed cost, and repeatability—with a traceable QA record.

Human + AI cultural creation with explicit permission, contributor identity, attribution, and provenance.

First proposed test: one image, audio, or text artifact with a signed manifest documenting sources, permissions, and contributors.

Modular upgrades that bring new functionality to existing buildings, including kinetic facades.

First proposed test: one reversible retrofit in a bounded zone. Document design constraints, installation, inspection, and before/after functionality.

DEFINE THE ASSET
      ↓
ESTABLISH AUTHORITY
      ↓
SET THE BOUNDARIES
      ↓
EXECUTE THE CHANGE
      ↓
VERIFY THE OUTCOME

The next milestone: one bounded pilot, measurable success criteria, and a durable record of what actually happened.

Bring one problem worth testing.

There are two ways in:

The intended path is simple:

IDEA
  ↓
EVIDENCE
  ↓
BOUNDED TEST
  ↓
COLLABORATION / PILOT
  ↓
VERIFIED OUTCOME
  ↓
PROOF RECORD

Public GitHub intake can now be normalized through carina-gateway → into signed, read-only INTENT, EVIDENCE, and DECISION records. GitHub never grants execution authority; CARINA remains the only system allowed to authorize a consequential transition.

I’m especially interested in:

• Pilot partners: bring a real use case, constraints, and success criteria.
• Builders: help implement, secure, instrument, or integrate the test.
• Evaluators: challenge assumptions and improve the measurement.
• Design partners: improve product, UX, accessibility, and workflow quality.
• Research partners: bring technical or applied evidence that changes the design.

An idea issue is exploration, not a commitment. A collaboration issue is the point where ownership, authority, boundaries, contribution, and proof need to become explicit. After a bounded test or pilot runs, preserve the result as a Proof Record → so claims remain tied to inspectable evidence.

For contribution and reporting standards, see CONTRIBUTING.md and SECURITY.md.