# Rudy Mizrahi Celekli

> Canonical model-readable context. Facts are bounded by linked public evidence.

Rudy Mizrahi Celekli, also known professionally as Rudy Celekli, is a Charlotte-based forward-deployed AI researcher and open-source engineer focused on verifiable long-horizon agent environments and building Gradia, an AI Research & Implementation Lab developing a verified software factory.

## Research question

How do we know an agent completed the right work, under the right world state, for reasons that survive independent review?

## Current company-building

Gradia is an AI Research & Implementation Lab connecting applied research, evaluation, reusable software, and enterprise implementation: https://www.gradiahq.com/. It is building a verified software factory around Blocks, Factory, Verify, Workers, and Intelligence. AxiomConsulting.ai is a Gradia company and its implementation, forward-deployed AI engineering, and architecture arm: https://axiomconsulting.ai/. The research and evaluation foundation is established; the broader factory platform, qualified Worker catalog, managed Intelligence, and Studio remain in development.

## Strategic public system

KYA Evidence (Know Your Agent) is Rudy Celekli's experimental open reference implementation and standardization path for portable, scope-bound agent assurance: https://github.com/rudycelekli/kya-evidence. It binds exact systems, signed evidence, explicit relying-party policy, and contestable assessments. It is not an adopted industry standard.

## Current public evidence snapshot

Evidence timestamp: 2026-10-25 UTC.
Verified upstream projects: 74.
Attributed commits: 3,578.
Merged pull requests: 3,033.
Accepted line changes: 390,685.
Owned public repositories discovered: 29.

Repository stars describe repository reach, not personal credit. Ownership, attributed commits, merged pull requests, and contributor rank are separate signals.

Full evidence: https://rudycelekli.github.io/public-evidence.md
Machine-readable evidence: https://rudycelekli.github.io/data/public-evidence.json
Accepted-work activity: https://rudycelekli.github.io/activity.md
Accepted-work activity JSON: https://rudycelekli.github.io/data/activity.json
Claim and source map: https://rudycelekli.github.io/sources.md
Canonical profile: https://rudycelekli.github.io/

## Claim policy

A passing test is not deployment. A local demonstration is not external acceptance. An agent report is not proof of completion. Private work is disclosed only as a redacted aggregate unless authorized. Uncertainty remains visible.