About
Ariamena exists because useful AI cannot be trained on abstraction alone. It needs real human context, real environments, responsible collection, and rigorous structure.
Founders
Three people who each watched AI fail in an environment they knew well, and decided the data was the problem.
Omar El-Sayed
Co-founder, Chief Executive
Spent a decade running operations in manufacturing and logistics before moving into AI. Started Ariamena after watching models trained on clean lab data fail on real floors.
- Program design
- Partner sites
- Commercial
Karim Haddad
Co-founder, Chief Technology
Built data pipelines and evaluation tooling for computer vision and speech teams. Owns the capture stack, annotation tooling, and delivery formats.
- Capture and annotation stack
- Data quality systems
- Delivery
Layla Mansour
Co-founder, Head of Programs and Responsible Data
Field researcher by training, with years of consent-based fieldwork in homes, clinics, and schools. Designs how contributors are briefed, protected, and credited.
- Contributor network
- Consent and governance
- Field operations
The name
Ariamena joins two ideas. Aria: a single human voice, given form and heard clearly. Mena: the people and places the company comes from. Together they describe the work: human knowledge, made legible for intelligent systems, without losing the people behind it.
What we believe
Context is the product
A recording without its setting, sequence, and intent is not yet useful. We capture the context with the signal.
People are partners
Contributors are briefed, consented, trained, and credited in the program's records. They are not a crowd.
Rigor is a form of respect
Acceptance criteria, review, and traceability protect the people behind the data as much as the model built from it.
Quiet claims, clear evidence
We would rather show a method than announce a number.
Work with us
Ariamena was founded in 2025 by three people who had each watched AI fail in environments they knew well. We are building a team around that experience. We want to hear from people who know real environments from the inside, from operations leaders who want AI that understands their work, and from research and engineering partners who want a data partner rather than a vendor.
Tell us what your AI needs to understand. We'll help shape the data program that gets it there.