Skip to main content

Ethics and responsibility

Five principles that govern how we build.

Systems that reflect the best of humanity.

Trustworthy autonomous AI, built in Latin America for operations where context, safety, and measurable results matter.

  1. 01

    VERIFIABLE TRUST

    We build autonomous AI systems that can be observed, audited, and stopped. Guardrails, action logs, and human approval are part of the design, not a patch added after launch.

  2. 02

    AI THAT REMOVES WORK

    We automate the repetitive, high-friction parts of operations so teams spend their time on judgment, not busywork. We prioritize the people technology usually leaves behind: women and workers in Mexico and Latin America.

  3. 03

    PROFITABLE, WITH CLEAR LINES

    We are building a company that grows and earns. Our limits are public: no direct defense work, and persuasion only with consent, never manipulation.

AI should serve humanity, not the reverse.

The five principles

Our operating protocol

  1. 01

    TECHNICAL SOVEREIGNTY

    Local models first, cloud only when it adds value. Client data stays under the client's control, and we run capable open models on modest hardware to keep costs predictable.

  2. 02

    RIGOR + CREATIVITY

    Decisions rest on mathematical analysis (UNAM), production infrastructure experience (AWS, Santander), and measurable tests. Creativity picks what is worth building. Rigor decides when it is ready to ship.

  3. 03

    BUILT FROM REAL FRICTION

    Every product starts from a problem we can measure: job search in Mexico is fragmented (Ergane), and the pay gap is quantifiable (FairHire). No vanity projects.

  4. 04

    DEMOCRATIZE, DON'T CONCENTRATE

    AI should widen access, not concentrate power in a few companies. We build tools that small teams and independent workers in Latin America can afford and operate.

  5. 05

    RADICAL TRANSPARENCY

    We share selected repositories and public datasets, including HuggingFace `gio5464/benchmarks`. Our product repositories remain private.

Building beyond the horizon

Each principle comes from a real problem we hit building for Mexico and Latin America. If you need an AI partner you can audit, let's talk.